yolo_512x512_8_working_better.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "igEd7obtcxR2"
   },
   "source": [
    "## Approach\n",
    "\n",
    "* This is a simplied version of YOLO. We take our 512x512 input, divide it into a 4x4 grid, each cell outputs five values:\n",
    "    * The confidence that there is pneumonia present\n",
    "    * The x, y, w, and h of the bounding box\n",
    "    * We have removed the entire classification section, using the confidence to indicate whether there is an ROI in the cell instead.\n",
    "* A sigmoid is applied to the output of the network to result in values between 0 and 1    \n",
    "* The x and y coordinates are offsets from the upper left corner of each cell, the w and h are percentage of the total width.\n",
    "* The loss function is based on YOLO with some differences:\n",
    "    * The weights of the components have been updated\n",
    "    * The \"objectness\" loss of YOLO tries to make the confidence match the actual IOU. This pushed the confidence of cells with low IOUs down. Since we only have one box per cell, this doesn't make any sense so has been removed. We have replaced it by removing the confidence and simply subtracting the IOU of cells with ROIs from 1 to maximize the IOU.\n",
    "\n",
    "## Network\n",
    "\n",
    "* The network consists of a number of residual blocks with convolutions and downsampling blocks with max pooling.\n",
    "* The network outputs a 4x4x5 array as described above.\n",
    "\n",
    "## Predictions\n",
    "* To generate our predictions we loop through each cell of the output\n",
    "* We unnormalize the output to get the actual values\n",
    "* If the confidence is greater than 0.5 and the box has width and height we append it to the list of candidates.\n",
    "* We apply non-max suppression to remove duplicates.\n",
    "* The remaining boxes are concatenated onto the output string.\n",
    "\n",
    "**Change Log:**\n",
    "* v3 - changing output to 8x8 grid from 16x16; changed model to downsample one more time; adjusted network accordingly. \n",
    "* v4 - changed output to 4x4 grid, no image has more than 3 ROIs so this may work better? \n",
    "    * Using center point of ROI to predict instead of upper left corner.\n",
    "* v5 - We only calculate MSE loss for boxes with a confidence over 0.5 or actual truth since we don't care about predictions for boxes that are not ROIs. This will prevent the network from being constrained by outputting 0s for boxes that don't exist.\n",
    "* v6 - centering input data so maybe bboxes can be output more accurately? Also centering the image\n",
    "* v6.2 - labels have ROI centered in center of cell by default instead of mean location.\n",
    "* v8 - using custom loss function based on YOLO loss. Set default height and width to 1 px because 0 sent the gradients to -inf which screwed everything up."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "LHncOor-cxSS"
   },
   "outputs": [],
   "source": [
    "import os\n",
    "import csv\n",
    "import random\n",
    "import pydicom\n",
    "import numpy as np\n",
    "import pickle\n",
    "import pandas as pd\n",
    "from skimage import measure\n",
    "from skimage.transform import resize\n",
    "import datetime\n",
    "import math\n",
    "\n",
    "import tensorflow as tf\n",
    "from tensorflow import keras\n",
    "\n",
    "from matplotlib import pyplot as plt\n",
    "import matplotlib.patches as patches"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Malisiewicz et al.\n",
    "def non_max_suppression_fast(boxes, overlapThresh=0.3):\n",
    "    # if there are no boxes, return an empty list\n",
    "    if len(boxes) == 0:\n",
    "        return []\n",
    " \n",
    "    # if the bounding boxes integers, convert them to floats --\n",
    "    # this is important since we'll be doing a bunch of divisions\n",
    "    if boxes.dtype.kind == \"i\":\n",
    "        boxes = boxes.astype(\"float\")\n",
    " \n",
    "    # initialize the list of picked indexes\t\n",
    "    pick = []\n",
    " \n",
    "    # grab the coordinates of the bounding boxes\n",
    "    x1 = boxes[:,0]\n",
    "    y1 = boxes[:,1]\n",
    "    w = boxes[:,2]\n",
    "    h = boxes[:,3]\n",
    " \n",
    "    x2 = x1 + w\n",
    "    y2 = y1 + h\n",
    "\n",
    "    # compute the area of the bounding boxes and sort the bounding\n",
    "    # boxes by the bottom-right y-coordinate of the bounding box\n",
    "    area = (w + 1) * (h + 1)\n",
    "    idxs = np.argsort(y2)\n",
    " \n",
    "    # keep looping while some indexes still remain in the indexes\n",
    "    # list\n",
    "    while len(idxs) > 0:\n",
    "        # grab the last index in the indexes list and add the\n",
    "        # index value to the list of picked indexes\n",
    "        last = len(idxs) - 1\n",
    "        i = idxs[last]\n",
    "        pick.append(i)\n",
    " \n",
    "        # find the largest (x, y) coordinates for the start of\n",
    "        # the bounding box and the smallest (x, y) coordinates\n",
    "        # for the end of the bounding box\n",
    "        xx1 = np.maximum(x1[i], x1[idxs[:last]])\n",
    "        yy1 = np.maximum(y1[i], y1[idxs[:last]])\n",
    "        xx2 = np.minimum(x2[i], x2[idxs[:last]])\n",
    "        yy2 = np.minimum(y2[i], y2[idxs[:last]])\n",
    " \n",
    "        # compute the width and height of the bounding box\n",
    "        w = np.maximum(0, xx2 - xx1 + 1)\n",
    "        h = np.maximum(0, yy2 - yy1 + 1)\n",
    " \n",
    "        # compute the ratio of overlap\n",
    "        overlap = (w * h) / area[idxs[:last]]\n",
    " \n",
    "        # delete all indexes from the index list that have\n",
    "        idxs = np.delete(idxs, np.concatenate(([last],\n",
    "            np.where(overlap > overlapThresh)[0])))\n",
    " \n",
    "    # return only the bounding boxes that were picked using the\n",
    "    # integer data type\n",
    "    return boxes[pick].astype(\"int\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "hE3jDmgudDDP"
   },
   "outputs": [],
   "source": [
    "# enter your Kaggle credentionals here\n",
    "os.environ['KAGGLE_USERNAME']=\"skooch\"\n",
    "os.environ['KAGGLE_KEY']=\"42f8a02ee92cc773d1dbe66565673ad3\""
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "def sigmoid(x):\n",
    "    y = np.maximum(x, -700)\n",
    "    return 1 / (1 + np.exp(-y))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "sSWY-U2TcxSn"
   },
   "source": [
    "# Load pneumonia locations\n",
    "\n",
    "Table contains [filename : pneumonia location] pairs per row. \n",
    "* If a filename contains multiple pneumonia, the table contains multiple rows with the same filename but different pneumonia locations. \n",
    "* If a filename contains no pneumonia it contains a single row with an empty pneumonia location.\n",
    "\n",
    "The code below loads the table and transforms it into a dictionary. \n",
    "* The dictionary uses the filename as key and a list of pneumonia locations in that filename as value. \n",
    "* If a filename is not present in the dictionary it means that it contains no pneumonia."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 34
    },
    "colab_type": "code",
    "id": "m_0ZGQyLdQS8",
    "outputId": "4848c0ce-4d75-4781-d01f-91cbd7626d77"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Train dir: ./stage_1_train_images\n"
     ]
    }
   ],
   "source": [
    "ROOT_DIR = \"./\"\n",
    "\n",
    "train_dicom_dir = os.path.join(ROOT_DIR, 'stage_1_train_images')\n",
    "test_dicom_dir = os.path.join(ROOT_DIR, 'stage_1_test_images')\n",
    "print(\"Train dir:\", train_dicom_dir)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "hopsRmICcxSs"
   },
   "outputs": [],
   "source": [
    "with open('yolo_labels_centered_4x4_6.p', 'rb') as handle:\n",
    "    pneumonia_locations = pickle.load(handle)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "UoB6GBhIcxS6"
   },
   "source": [
    "# Load filenames"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 51
    },
    "colab_type": "code",
    "id": "b_iB-IftcxTE",
    "outputId": "89a05e05-f2d9-45fc-d1ef-1a0cc289f98d"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "n train samples 23116\n",
      "n valid samples 2568\n"
     ]
    }
   ],
   "source": [
    "random.seed(72)\n",
    "\n",
    "# load and shuffle filenames\n",
    "folder = './stage_1_train_images'\n",
    "filenames = os.listdir(folder)\n",
    "random.shuffle(filenames)\n",
    "# split into train and validation filenames\n",
    "n_valid_samples = int(len(filenames) * 0.1)\n",
    "train_filenames = filenames[n_valid_samples:]\n",
    "valid_filenames = filenames[:n_valid_samples]\n",
    "print('n train samples', len(train_filenames))\n",
    "print('n valid samples', len(valid_filenames))\n",
    "n_train_samples = len(filenames) - n_valid_samples"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "lorV7ZmDcxTa"
   },
   "source": [
    " # Data generator\n",
    "\n",
    "The dataset is too large to fit into memory, so we need to create a generator that loads data on the fly.\n",
    "\n",
    "* The generator takes in some filenames, batch_size and other parameters.\n",
    "\n",
    "* The generator outputs a random batch of numpy images and numpy masks.\n",
    "    "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "dFbLdFxVehhB"
   },
   "outputs": [],
   "source": [
    "BATCH_SIZE = 14\n",
    "IMAGE_SIZE = 512\n",
    "CHECKPOINT_PATH = \"yolo8_2_512.h5\""
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [],
   "source": [
    "# means to center data\n",
    "mu_x = 391.456158\n",
    "mu_y = 363.1358768\n",
    "mu_w = 220.8453815\n",
    "mu_h = 334.1743641\n",
    "mu_center_x = 501.8788487\n",
    "mu_center_y = 530.2230589"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "TueY1bVlcxTg"
   },
   "outputs": [],
   "source": [
    "class generator(keras.utils.Sequence):\n",
    "    \n",
    "    def __init__(self, folder, filenames, pneumonia_locations=None, batch_size=BATCH_SIZE, image_size=IMAGE_SIZE, shuffle=True, augment=False, predict=False):\n",
    "        self.folder = folder\n",
    "        self.filenames = filenames\n",
    "        self.pneumonia_locations = pneumonia_locations\n",
    "        self.batch_size = batch_size\n",
    "        self.image_size = image_size\n",
    "        self.shuffle = shuffle\n",
    "        self.augment = augment\n",
    "        self.predict = predict\n",
    "        self.on_epoch_end()\n",
    "        \n",
    "    def __load__(self, filename):\n",
    "        # load dicom file as numpy array\n",
    "        img = pydicom.dcmread(os.path.join(self.folder, filename)).pixel_array\n",
    "        # get filename without extension\n",
    "        filename = filename.split('.')[0]\n",
    "        label = pneumonia_locations[filename].copy()\n",
    "        \n",
    "        # remove the confidence and bboxes because they will be flipped separately\n",
    "        confs = label[:,:,0]\n",
    "        boxes = label[:,:,1:]\n",
    "        \n",
    "        ## data augmentation may be complicated, let's do that later\n",
    "        # if augment then horizontal flip half the time\n",
    "        if self.augment and random.random() > 0.5:\n",
    "            # flip the image\n",
    "            img = np.fliplr(img)\n",
    "            \n",
    "            # update our x coords\n",
    "            mask = confs != 0\n",
    "            \n",
    "            # flip\n",
    "            boxes[mask, 0] = 1 - boxes[mask,0]\n",
    "            \n",
    "            # flip our boxes lr on axis 0\n",
    "            boxes = np.flip(boxes, axis=0)         \n",
    "            \n",
    "            # flip the confidences lr as well\n",
    "            confs = np.flip(confs, axis=0)\n",
    "            \n",
    "        # resize both image and mask\n",
    "        img = resize(img, (self.image_size, self.image_size), mode='reflect')\n",
    "        \n",
    "        # center the image\n",
    "        img = img - 0.5\n",
    "        \n",
    "        boxes = np.concatenate([confs.reshape((4,4,1)), boxes], axis=2)\n",
    "        \n",
    "        # add trailing channel dimension\n",
    "        img = np.expand_dims(img, -1)\n",
    "        return img, boxes\n",
    "    \n",
    "    def __loadpredict__(self, filename):\n",
    "        # load dicom file as numpy array\n",
    "        img = pydicom.dcmread(os.path.join(self.folder, filename)).pixel_array\n",
    "        # resize image\n",
    "        img = resize(img, (self.image_size, self.image_size), mode='reflect')\n",
    "        # add trailing channel dimension\n",
    "        img = np.expand_dims(img, -1)\n",
    "        return img\n",
    "        \n",
    "    def __getitem__(self, index):\n",
    "        # select batch\n",
    "        filenames = self.filenames[index*self.batch_size:(index+1)*self.batch_size]\n",
    "        # predict mode: return images and filenames\n",
    "        if self.predict:\n",
    "            # load files\n",
    "            imgs = [self.__loadpredict__(filename) for filename in filenames]\n",
    "            # create numpy batch\n",
    "            imgs = np.array(imgs)\n",
    "            return imgs, filenames\n",
    "        # train mode: return images and masks\n",
    "        else:\n",
    "            # load files\n",
    "            items = [self.__load__(filename) for filename in filenames]\n",
    "            \n",
    "            # unzip images and masks\n",
    "            imgs, bboxes = zip(*items)\n",
    "            \n",
    "            # create numpy batch\n",
    "            imgs = np.array(imgs)\n",
    "            bboxes = np.array(bboxes)\n",
    "            \n",
    "            labels = bboxes\n",
    "            return imgs, labels\n",
    "        \n",
    "    def on_epoch_end(self):\n",
    "        if self.shuffle:\n",
    "            random.shuffle(self.filenames)\n",
    "        \n",
    "    def __len__(self):\n",
    "        if self.predict:\n",
    "            # return everything\n",
    "            return int(np.ceil(len(self.filenames) / self.batch_size))\n",
    "        else:\n",
    "            # return full batches only\n",
    "            return int(len(self.filenames) / self.batch_size)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/home/eric/.local/lib/python3.5/site-packages/skimage/transform/_warps.py:110: UserWarning: Anti-aliasing will be enabled by default in skimage 0.15 to avoid aliasing artifacts when down-sampling images.\n",
      "  warn(\"Anti-aliasing will be enabled by default in skimage 0.15 to \"\n"
     ]
    }
   ],
   "source": [
    "train_gen = generator(folder, train_filenames, pneumonia_locations, batch_size=BATCH_SIZE, image_size=IMAGE_SIZE, shuffle=True, augment=True, predict=False)\n",
    "\n",
    "counter = 0\n",
    "for imgs, labels in train_gen:\n",
    "    for label in labels:\n",
    "        for i in range(4):\n",
    "            for j in range(4):\n",
    "                if label[i,j,0] == 1:\n",
    "                    if label[i,j,4] ==  0.001:\n",
    "                        print(\"Error!\")\n",
    "    counter += 1\n",
    "    if counter > 15:\n",
    "        break"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "3_GgDL7ncxT3"
   },
   "source": [
    "# Network"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "bqobZQZQcxUE"
   },
   "outputs": [],
   "source": [
    "def create_downsample(channels, inputs):\n",
    "    x = keras.layers.BatchNormalization(momentum=0.9)(inputs)\n",
    "    x = keras.layers.LeakyReLU(0.1)(x)\n",
    "    x = keras.layers.MaxPool2D(2)(x)\n",
    "    return x\n",
    "\n",
    "def create_resblock(channels, inputs):\n",
    "    x = keras.layers.BatchNormalization(momentum=0.9)(inputs)\n",
    "    x = keras.layers.LeakyReLU(0.1)(x)\n",
    "    x_1 = keras.layers.Conv2D(channels, 3, padding='same', use_bias=False)(x)\n",
    "    x = keras.layers.BatchNormalization(momentum=0.9)(x_1)\n",
    "    x = keras.layers.LeakyReLU(0.1)(x)\n",
    "    x = keras.layers.Conv2D(channels, 3, padding='same', use_bias=False)(x)\n",
    "    x = keras.layers.BatchNormalization(momentum=0.9)(x)\n",
    "    x = keras.layers.LeakyReLU(0.1)(x)\n",
    "    x = keras.layers.Conv2D(channels, 3, padding='same', use_bias=False)(x)\n",
    "    x = keras.layers.BatchNormalization(momentum=0.9)(x)\n",
    "    x = keras.layers.LeakyReLU(0.1)(x)\n",
    "    x = keras.layers.Conv2D(channels, 3, padding='same', use_bias=False)(x)\n",
    "    x = keras.layers.add([x, x_1])\n",
    "    return x\n",
    "\n",
    "def create_network(input_size, channels, n_blocks=2, depth=4):\n",
    "    # input - 512x512x3\n",
    "    inputs = keras.Input(shape=(input_size, input_size, 1))\n",
    "    \n",
    "    # 256x256x24\n",
    "    x = keras.layers.Conv2D(channels, 3, strides=(2,2), padding='same', use_bias=False)(inputs)\n",
    "    x = keras.layers.BatchNormalization(momentum=0.9)(x)\n",
    "    x = keras.layers.LeakyReLU(0.1)(x)\n",
    "    \n",
    "    x = keras.layers.Conv2D(channels, 3, padding='same', use_bias=False)(x)\n",
    "    x = keras.layers.BatchNormalization(momentum=0.9)(x)\n",
    "    x = keras.layers.LeakyReLU(0.1)(x)\n",
    "    \n",
    "    x = keras.layers.Conv2D(channels, 3, padding='same', use_bias=False)(x)\n",
    "    x = keras.layers.BatchNormalization(momentum=0.9)(x)\n",
    "    x = keras.layers.LeakyReLU(0.1)(x)\n",
    "    \n",
    "    # 256x256x24 for residual connection\n",
    "    x = keras.layers.Conv2D(channels, 1, padding='same', use_bias=False)(x)\n",
    "    \n",
    "    # residual blocks\n",
    "    for d in range(depth):\n",
    "        channels = channels * 2\n",
    "        \n",
    "        x = create_downsample(channels, x)\n",
    "            \n",
    "        for b in range(n_blocks):\n",
    "            x = create_resblock(channels, x)\n",
    "    \n",
    "    x = keras.layers.BatchNormalization(momentum=0.9)(x)\n",
    "    x = keras.layers.LeakyReLU(0)(x)\n",
    "    \n",
    "    # dilated convolutions for context - 16x16x512\n",
    "    x_2 = keras.layers.Conv2D(512, (3,3), padding='same', dilation_rate=(2,2), activation=None, name=\"dilated_conv_1\")(x)\n",
    "    x = keras.layers.BatchNormalization(momentum=0.9)(x_2)\n",
    "    x = keras.layers.LeakyReLU(0.1)(x)\n",
    "    \n",
    "    x = keras.layers.Conv2D(512, (3,3), padding='same', dilation_rate=(2,2), activation=None, name=\"dilated_conv_2\")(x)\n",
    "    x = keras.layers.BatchNormalization(momentum=0.9)(x)\n",
    "    x = keras.layers.LeakyReLU(0.1)(x)\n",
    "    \n",
    "    x = keras.layers.Conv2D(512, (3,3), padding='same', activation=None, name=\"last_conv\")(x)\n",
    "    x = keras.layers.add([x, x_2])\n",
    "    x = keras.layers.BatchNormalization(momentum=0.9)(x)\n",
    "    x = keras.layers.LeakyReLU(0.1)(x)\n",
    "    \n",
    "    # downsample to 4x4x512 with stride 4\n",
    "    x = keras.layers.Conv2D(512, (4,4), padding='same', strides=(4,4), activation=None, name=\"downsample_1\")(x)\n",
    "    x = keras.layers.BatchNormalization(momentum=0.9)(x)\n",
    "    x = keras.layers.LeakyReLU(0.1)(x)\n",
    "    \n",
    "#     # confidence branch\n",
    "#     c = keras.layers.Conv2D(384, (1,1), padding='same', activation=None, name=\"fc_1_c\", kernel_regularizer=keras.regularizers.l2(l=0.01))(x)\n",
    "#     c = keras.layers.BatchNormalization(momentum=0.9)(c)\n",
    "#     c = keras.layers.LeakyReLU(0.1)(c)\n",
    "#     c = keras.layers.Dropout(0.25)(c)\n",
    "    \n",
    "#     c = keras.layers.Conv2D(256, (1,1), padding='same', activation=None, name=\"fc_2_c\", kernel_regularizer=keras.regularizers.l2(l=0.01))(c)\n",
    "#     c = keras.layers.BatchNormalization(momentum=0.9)(c)\n",
    "#     c = keras.layers.LeakyReLU(0.1)(c)\n",
    "#     c = keras.layers.Dropout(0.10)(c)\n",
    "    \n",
    "#     confidence = keras.layers.Conv2D(1, (1,1), padding='same', activation=None, name=\"confidence_output\")(c)\n",
    "    \n",
    "    # bounding box branch\n",
    "    b = keras.layers.Conv2D(1024, (1,1), padding='same', activation=None, name=\"fc_1_b\", kernel_regularizer=keras.regularizers.l2(l=0.005))(x)\n",
    "    b = keras.layers.BatchNormalization(momentum=0.9)(b)\n",
    "    b = keras.layers.LeakyReLU(0.01)(b)\n",
    "    b = keras.layers.Dropout(0.15)(b)\n",
    "    \n",
    "    b = keras.layers.Conv2D(1024, (1,1), padding='same', activation=None, name=\"fc_2_b\", kernel_regularizer=keras.regularizers.l2(l=0.001))(b)\n",
    "    b = keras.layers.LeakyReLU(0.01)(b)\n",
    "#     b = keras.layers.Dropout(0.05)(b)\n",
    "    \n",
    "    boxes = keras.layers.Conv2D(5, (1,1), padding='same', activation=\"linear\")(b)\n",
    "#     boxes = keras.layers.concatenate([confidence, boxes], name=\"bboxes_output\")\n",
    "    \n",
    "    # return both outputs\n",
    "    model = keras.Model(inputs=inputs, outputs=boxes)\n",
    "    \n",
    "    return model"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "AvdSzGI4cxUL"
   },
   "source": [
    "# Train network\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 3451
    },
    "colab_type": "code",
    "id": "DCgIz8CwcxUO",
    "outputId": "ce38cdfb-1806-488d-838f-d180e198a099"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "__________________________________________________________________________________________________\n",
      "Layer (type)                    Output Shape         Param #     Connected to                     \n",
      "==================================================================================================\n",
      "input_4 (InputLayer)            (None, 512, 512, 1)  0                                            \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_64 (Conv2D)              (None, 256, 256, 24) 216         input_4[0][0]                    \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_88 (BatchNo (None, 256, 256, 24) 96          conv2d_64[0][0]                  \n",
      "__________________________________________________________________________________________________\n",
      "leaky_re_lu_91 (LeakyReLU)      (None, 256, 256, 24) 0           batch_normalization_88[0][0]     \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_65 (Conv2D)              (None, 256, 256, 24) 5184        leaky_re_lu_91[0][0]             \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_89 (BatchNo (None, 256, 256, 24) 96          conv2d_65[0][0]                  \n",
      "__________________________________________________________________________________________________\n",
      "leaky_re_lu_92 (LeakyReLU)      (None, 256, 256, 24) 0           batch_normalization_89[0][0]     \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_66 (Conv2D)              (None, 256, 256, 24) 5184        leaky_re_lu_92[0][0]             \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_90 (BatchNo (None, 256, 256, 24) 96          conv2d_66[0][0]                  \n",
      "__________________________________________________________________________________________________\n",
      "leaky_re_lu_93 (LeakyReLU)      (None, 256, 256, 24) 0           batch_normalization_90[0][0]     \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_67 (Conv2D)              (None, 256, 256, 24) 576         leaky_re_lu_93[0][0]             \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_91 (BatchNo (None, 256, 256, 24) 96          conv2d_67[0][0]                  \n",
      "__________________________________________________________________________________________________\n",
      "leaky_re_lu_94 (LeakyReLU)      (None, 256, 256, 24) 0           batch_normalization_91[0][0]     \n",
      "__________________________________________________________________________________________________\n",
      "max_pooling2d_13 (MaxPooling2D) (None, 128, 128, 24) 0           leaky_re_lu_94[0][0]             \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_92 (BatchNo (None, 128, 128, 24) 96          max_pooling2d_13[0][0]           \n",
      "__________________________________________________________________________________________________\n",
      "leaky_re_lu_95 (LeakyReLU)      (None, 128, 128, 24) 0           batch_normalization_92[0][0]     \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_68 (Conv2D)              (None, 128, 128, 48) 10368       leaky_re_lu_95[0][0]             \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_93 (BatchNo (None, 128, 128, 48) 192         conv2d_68[0][0]                  \n",
      "__________________________________________________________________________________________________\n",
      "leaky_re_lu_96 (LeakyReLU)      (None, 128, 128, 48) 0           batch_normalization_93[0][0]     \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_69 (Conv2D)              (None, 128, 128, 48) 20736       leaky_re_lu_96[0][0]             \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_94 (BatchNo (None, 128, 128, 48) 192         conv2d_69[0][0]                  \n",
      "__________________________________________________________________________________________________\n",
      "leaky_re_lu_97 (LeakyReLU)      (None, 128, 128, 48) 0           batch_normalization_94[0][0]     \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_70 (Conv2D)              (None, 128, 128, 48) 20736       leaky_re_lu_97[0][0]             \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_95 (BatchNo (None, 128, 128, 48) 192         conv2d_70[0][0]                  \n",
      "__________________________________________________________________________________________________\n",
      "leaky_re_lu_98 (LeakyReLU)      (None, 128, 128, 48) 0           batch_normalization_95[0][0]     \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_71 (Conv2D)              (None, 128, 128, 48) 20736       leaky_re_lu_98[0][0]             \n",
      "__________________________________________________________________________________________________\n",
      "add_16 (Add)                    (None, 128, 128, 48) 0           conv2d_71[0][0]                  \n",
      "                                                                 conv2d_68[0][0]                  \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_96 (BatchNo (None, 128, 128, 48) 192         add_16[0][0]                     \n",
      "__________________________________________________________________________________________________\n",
      "leaky_re_lu_99 (LeakyReLU)      (None, 128, 128, 48) 0           batch_normalization_96[0][0]     \n",
      "__________________________________________________________________________________________________\n",
      "max_pooling2d_14 (MaxPooling2D) (None, 64, 64, 48)   0           leaky_re_lu_99[0][0]             \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_97 (BatchNo (None, 64, 64, 48)   192         max_pooling2d_14[0][0]           \n",
      "__________________________________________________________________________________________________\n",
      "leaky_re_lu_100 (LeakyReLU)     (None, 64, 64, 48)   0           batch_normalization_97[0][0]     \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_72 (Conv2D)              (None, 64, 64, 96)   41472       leaky_re_lu_100[0][0]            \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_98 (BatchNo (None, 64, 64, 96)   384         conv2d_72[0][0]                  \n",
      "__________________________________________________________________________________________________\n",
      "leaky_re_lu_101 (LeakyReLU)     (None, 64, 64, 96)   0           batch_normalization_98[0][0]     \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_73 (Conv2D)              (None, 64, 64, 96)   82944       leaky_re_lu_101[0][0]            \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_99 (BatchNo (None, 64, 64, 96)   384         conv2d_73[0][0]                  \n",
      "__________________________________________________________________________________________________\n",
      "leaky_re_lu_102 (LeakyReLU)     (None, 64, 64, 96)   0           batch_normalization_99[0][0]     \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_74 (Conv2D)              (None, 64, 64, 96)   82944       leaky_re_lu_102[0][0]            \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_100 (BatchN (None, 64, 64, 96)   384         conv2d_74[0][0]                  \n",
      "__________________________________________________________________________________________________\n",
      "leaky_re_lu_103 (LeakyReLU)     (None, 64, 64, 96)   0           batch_normalization_100[0][0]    \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_75 (Conv2D)              (None, 64, 64, 96)   82944       leaky_re_lu_103[0][0]            \n",
      "__________________________________________________________________________________________________\n",
      "add_17 (Add)                    (None, 64, 64, 96)   0           conv2d_75[0][0]                  \n",
      "                                                                 conv2d_72[0][0]                  \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_101 (BatchN (None, 64, 64, 96)   384         add_17[0][0]                     \n",
      "__________________________________________________________________________________________________\n",
      "leaky_re_lu_104 (LeakyReLU)     (None, 64, 64, 96)   0           batch_normalization_101[0][0]    \n",
      "__________________________________________________________________________________________________\n",
      "max_pooling2d_15 (MaxPooling2D) (None, 32, 32, 96)   0           leaky_re_lu_104[0][0]            \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_102 (BatchN (None, 32, 32, 96)   384         max_pooling2d_15[0][0]           \n",
      "__________________________________________________________________________________________________\n",
      "leaky_re_lu_105 (LeakyReLU)     (None, 32, 32, 96)   0           batch_normalization_102[0][0]    \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_76 (Conv2D)              (None, 32, 32, 192)  165888      leaky_re_lu_105[0][0]            \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_103 (BatchN (None, 32, 32, 192)  768         conv2d_76[0][0]                  \n",
      "__________________________________________________________________________________________________\n",
      "leaky_re_lu_106 (LeakyReLU)     (None, 32, 32, 192)  0           batch_normalization_103[0][0]    \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_77 (Conv2D)              (None, 32, 32, 192)  331776      leaky_re_lu_106[0][0]            \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_104 (BatchN (None, 32, 32, 192)  768         conv2d_77[0][0]                  \n",
      "__________________________________________________________________________________________________\n",
      "leaky_re_lu_107 (LeakyReLU)     (None, 32, 32, 192)  0           batch_normalization_104[0][0]    \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_78 (Conv2D)              (None, 32, 32, 192)  331776      leaky_re_lu_107[0][0]            \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_105 (BatchN (None, 32, 32, 192)  768         conv2d_78[0][0]                  \n",
      "__________________________________________________________________________________________________\n",
      "leaky_re_lu_108 (LeakyReLU)     (None, 32, 32, 192)  0           batch_normalization_105[0][0]    \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_79 (Conv2D)              (None, 32, 32, 192)  331776      leaky_re_lu_108[0][0]            \n",
      "__________________________________________________________________________________________________\n",
      "add_18 (Add)                    (None, 32, 32, 192)  0           conv2d_79[0][0]                  \n",
      "                                                                 conv2d_76[0][0]                  \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_106 (BatchN (None, 32, 32, 192)  768         add_18[0][0]                     \n",
      "__________________________________________________________________________________________________\n",
      "leaky_re_lu_109 (LeakyReLU)     (None, 32, 32, 192)  0           batch_normalization_106[0][0]    \n",
      "__________________________________________________________________________________________________\n",
      "max_pooling2d_16 (MaxPooling2D) (None, 16, 16, 192)  0           leaky_re_lu_109[0][0]            \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_107 (BatchN (None, 16, 16, 192)  768         max_pooling2d_16[0][0]           \n",
      "__________________________________________________________________________________________________\n",
      "leaky_re_lu_110 (LeakyReLU)     (None, 16, 16, 192)  0           batch_normalization_107[0][0]    \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_80 (Conv2D)              (None, 16, 16, 384)  663552      leaky_re_lu_110[0][0]            \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_108 (BatchN (None, 16, 16, 384)  1536        conv2d_80[0][0]                  \n",
      "__________________________________________________________________________________________________\n",
      "leaky_re_lu_111 (LeakyReLU)     (None, 16, 16, 384)  0           batch_normalization_108[0][0]    \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_81 (Conv2D)              (None, 16, 16, 384)  1327104     leaky_re_lu_111[0][0]            \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_109 (BatchN (None, 16, 16, 384)  1536        conv2d_81[0][0]                  \n",
      "__________________________________________________________________________________________________\n",
      "leaky_re_lu_112 (LeakyReLU)     (None, 16, 16, 384)  0           batch_normalization_109[0][0]    \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_82 (Conv2D)              (None, 16, 16, 384)  1327104     leaky_re_lu_112[0][0]            \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_110 (BatchN (None, 16, 16, 384)  1536        conv2d_82[0][0]                  \n",
      "__________________________________________________________________________________________________\n",
      "leaky_re_lu_113 (LeakyReLU)     (None, 16, 16, 384)  0           batch_normalization_110[0][0]    \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_83 (Conv2D)              (None, 16, 16, 384)  1327104     leaky_re_lu_113[0][0]            \n",
      "__________________________________________________________________________________________________\n",
      "add_19 (Add)                    (None, 16, 16, 384)  0           conv2d_83[0][0]                  \n",
      "                                                                 conv2d_80[0][0]                  \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_111 (BatchN (None, 16, 16, 384)  1536        add_19[0][0]                     \n",
      "__________________________________________________________________________________________________\n",
      "leaky_re_lu_114 (LeakyReLU)     (None, 16, 16, 384)  0           batch_normalization_111[0][0]    \n",
      "__________________________________________________________________________________________________\n",
      "dilated_conv_1 (Conv2D)         (None, 16, 16, 512)  1769984     leaky_re_lu_114[0][0]            \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_112 (BatchN (None, 16, 16, 512)  2048        dilated_conv_1[0][0]             \n",
      "__________________________________________________________________________________________________\n",
      "leaky_re_lu_115 (LeakyReLU)     (None, 16, 16, 512)  0           batch_normalization_112[0][0]    \n",
      "__________________________________________________________________________________________________\n",
      "dilated_conv_2 (Conv2D)         (None, 16, 16, 512)  2359808     leaky_re_lu_115[0][0]            \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_113 (BatchN (None, 16, 16, 512)  2048        dilated_conv_2[0][0]             \n",
      "__________________________________________________________________________________________________\n",
      "leaky_re_lu_116 (LeakyReLU)     (None, 16, 16, 512)  0           batch_normalization_113[0][0]    \n",
      "__________________________________________________________________________________________________\n",
      "last_conv (Conv2D)              (None, 16, 16, 512)  2359808     leaky_re_lu_116[0][0]            \n",
      "__________________________________________________________________________________________________\n",
      "add_20 (Add)                    (None, 16, 16, 512)  0           last_conv[0][0]                  \n",
      "                                                                 dilated_conv_1[0][0]             \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_114 (BatchN (None, 16, 16, 512)  2048        add_20[0][0]                     \n",
      "__________________________________________________________________________________________________\n",
      "leaky_re_lu_117 (LeakyReLU)     (None, 16, 16, 512)  0           batch_normalization_114[0][0]    \n",
      "__________________________________________________________________________________________________\n",
      "downsample_1 (Conv2D)           (None, 4, 4, 512)    4194816     leaky_re_lu_117[0][0]            \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_115 (BatchN (None, 4, 4, 512)    2048        downsample_1[0][0]               \n",
      "__________________________________________________________________________________________________\n",
      "leaky_re_lu_118 (LeakyReLU)     (None, 4, 4, 512)    0           batch_normalization_115[0][0]    \n",
      "__________________________________________________________________________________________________\n",
      "fc_1_b (Conv2D)                 (None, 4, 4, 1024)   525312      leaky_re_lu_118[0][0]            \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_116 (BatchN (None, 4, 4, 1024)   4096        fc_1_b[0][0]                     \n",
      "__________________________________________________________________________________________________\n",
      "leaky_re_lu_119 (LeakyReLU)     (None, 4, 4, 1024)   0           batch_normalization_116[0][0]    \n",
      "__________________________________________________________________________________________________\n",
      "dropout_4 (Dropout)             (None, 4, 4, 1024)   0           leaky_re_lu_119[0][0]            \n",
      "__________________________________________________________________________________________________\n",
      "fc_2_b (Conv2D)                 (None, 4, 4, 1024)   1049600     dropout_4[0][0]                  \n",
      "__________________________________________________________________________________________________\n",
      "leaky_re_lu_120 (LeakyReLU)     (None, 4, 4, 1024)   0           fc_2_b[0][0]                     \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_84 (Conv2D)              (None, 4, 4, 5)      5125        leaky_re_lu_120[0][0]            \n",
      "==================================================================================================\n",
      "Total params: 18,470,205\n",
      "Trainable params: 18,457,389\n",
      "Non-trainable params: 12,816\n",
      "__________________________________________________________________________________________________\n",
      "None\n"
     ]
    }
   ],
   "source": [
    "def binary_accuracy(y_true, y_pred):\n",
    "    y_pred = tf.sigmoid(y_pred)\n",
    "    y_true = tf.round(tf.reshape(y_true, [-1, 5]))\n",
    "    y_pred = tf.round(tf.reshape(y_pred, [-1, 5]))\n",
    "    \n",
    "    acc = tf.reduce_mean(tf.cast(tf.equal(y_true[:,0], y_pred[:,0]), dtype=tf.float32))\n",
    "    return acc\n",
    "    \n",
    "def overlap_iou(y_true, y_pred):\n",
    "    \"\"\"\n",
    "    Args:\n",
    "        bboxes1: shape (batch_size, 16, 16, 4)\n",
    "            with x1, y1, x2, y2 point order.\n",
    "        bboxes2: shape (batch_size, 16, 16, 4)\n",
    "            with x1, y1, x2, y2 point order.\n",
    "        p1 *-----\n",
    "           |     |\n",
    "           |_____* p2\n",
    "    Returns:\n",
    "        Tensor with shape (total_bboxes1, total_bboxes2)\n",
    "        with the IoU (intersection over union) of bboxes1[i] and bboxes2[j]\n",
    "        in [i, j].\n",
    "    \"\"\"\n",
    "    y_pred = tf.sigmoid(y_pred)\n",
    "    \n",
    "    # flatten the data because it's easier that way\n",
    "    bboxes1 = tf.reshape(y_true, (-1, 5))\n",
    "    bboxes2 = tf.reshape(y_pred, (-1, 5))\n",
    "    \n",
    "    # split the components out\n",
    "    true_boxes, x11, y11, w1, h1 = tf.split(bboxes1, 5, axis=1)\n",
    "    pred_conf, x21, y21, w2, h2 = tf.split(bboxes2, 5, axis=1)\n",
    "    \n",
    "    # uncenter the data - make sure the numbers are positive\n",
    "    x11 = x11 * 256\n",
    "    x21 = x21 * 256\n",
    "    y11 = y11 * 256\n",
    "    y21 = y21 * 256\n",
    "    \n",
    "    w1 = w1 * 1024\n",
    "    w2 = w2 * 1024\n",
    "    h1 = h1 * 1024\n",
    "    h2 = h2 * 1024\n",
    "    \n",
    "    # is a box predicted here? maybe use this for the IOU?\n",
    "    mask = (pred_conf >= 0.5) | (true_boxes == 1)\n",
    "    \n",
    "    # these labels are center points of the boxes so we need to calculate the corners\n",
    "    x12 = x11 + (w1 / 2)\n",
    "    y12 = y11 + (h1 / 2)\n",
    "    x22 = x21 + (w2 / 2)\n",
    "    y22 = y21 + (h2 / 2)\n",
    "    \n",
    "    x11 = x11 - (w1 / 2)\n",
    "    y11 = y11 - (h1 / 2)\n",
    "    x21 = x21 - (w2 / 2)\n",
    "    y21 = y21 - (h2 / 2)\n",
    "\n",
    "    # find the corners of the intersection area\n",
    "    xI1 = tf.maximum(x11, x21)\n",
    "    yI1 = tf.maximum(y11, y21)\n",
    "\n",
    "    xI2 = tf.minimum(x12, x22)\n",
    "    yI2 = tf.minimum(y12, y22)\n",
    "    \n",
    "    # get the intersection area, if the truth has no boxes it is 0\n",
    "    inter_area = true_boxes * (xI2 - xI1 + 1) * (yI2 - yI1 + 1)\n",
    "\n",
    "    # get the area of each box\n",
    "    bboxes1_area = (w1 + 1) * (h1 + 1)\n",
    "    bboxes2_area = (w2 + 1) * (h2 + 1)\n",
    "    \n",
    "    # union is area of both boxes - intersection\n",
    "    union = (bboxes1_area + bboxes2_area) - inter_area + 1\n",
    "    \n",
    "    # reduce the mean so we have mean iou for our inputs\n",
    "    return tf.reduce_mean(tf.maximum(inter_area / union, 0))\n",
    "\n",
    "def loss_fn(y_true, y_pred):\n",
    "    # get the iou loss\n",
    "    iou_loss = iou_loss_fn(y_true, y_pred)\n",
    "    \n",
    "    # get the xe loss\n",
    "    xe_loss = binary_cross_entropy(y_true, y_pred)\n",
    "    \n",
    "    # get the box loss\n",
    "    box_loss = adj_mse(y_true, y_pred)\n",
    "    \n",
    "    return xe_loss + (box_loss * 3.0) + (iou_loss * 2.0)\n",
    "\n",
    "# only apply mse to layers with high confidence that there is an ROI or if there actually is an ROI\n",
    "def adj_mse(y_true, y_pred):\n",
    "    # coefficients\n",
    "    lam_coord = 5\n",
    "    lam_noobj = 0.5\n",
    "    \n",
    "    y_pred = tf.sigmoid(y_pred)\n",
    "        \n",
    "    # flatten the inputs\n",
    "    y_true = tf.reshape(y_true, (-1, 5))\n",
    "    y_pred = tf.reshape(y_pred, (-1, 5))\n",
    "\n",
    "    # separate the confidence from the boxes\n",
    "    conf_true, x_true, y_true, w_true, h_true = tf.split(y_true, 5, axis=1)\n",
    "    conf_pred, x_pred, y_pred, w_pred, h_pred = tf.split(y_pred, 5, axis=1)\n",
    "\n",
    "    # center squared error\n",
    "    center_se = tf.square(x_true - y_pred) + tf.square(y_true - y_pred)\n",
    "    size_se =  tf.square(w_true - w_pred) + tf.square(h_true - h_pred)\n",
    "    \n",
    "    box_se = center_se + size_se\n",
    "    \n",
    "    # only get loss for boxes which are actually positive\n",
    "    mask = tf.cast(tf.equal(conf_true, 1), tf.float32) * 5.0\n",
    "    \n",
    "    box_se = tf.multiply(box_se, mask)\n",
    "    \n",
    "    loss = tf.reduce_sum(box_se) + 1e-16\n",
    "    \n",
    "    return loss\n",
    "    \n",
    "# use weight of 0.5 for negative cells, 19 for positive ones\n",
    "def binary_cross_entropy(y_true, y_pred):\n",
    "    conf_true = y_true[...,0]\n",
    "    conf_pred = tf.sigmoid(y_pred[...,0])\n",
    "    \n",
    "    weight = 10.0\n",
    "    weights = tf.multiply(y_true[...,0], weight) + 0.5\n",
    "    \n",
    "#     xe = tf.nn.sigmoid_cross_entropy_with_logits(logits=y_pred, labels=y_true)\n",
    "#     xe = tf.multiply(xe, weights)\n",
    "    \n",
    "    xe = tf.multiply(tf.square(conf_true - conf_pred), weights)\n",
    "    \n",
    "    return tf.reduce_sum(xe)\n",
    "\n",
    "def iou_loss_fn(y_true, y_pred):\n",
    "    # apply sigmoid to predictions\n",
    "    y_pred = tf.sigmoid(y_pred)\n",
    "    \n",
    "    # separate the x, y from the w, h and unnormalize them\n",
    "    true_box_xy = y_true[...,1:3] * 256\n",
    "    true_box_wh = y_true[...,3:] * 1024\n",
    "    \n",
    "    pred_box_xy = y_pred[...,1:3] * 256\n",
    "    pred_box_wh = y_pred[...,3:] * 1024\n",
    "    \n",
    "    # get the corners of the boxes by subtracting or adding half of h, w\n",
    "    true_wh_half = true_box_wh / 2.\n",
    "    true_mins    = true_box_xy - true_wh_half\n",
    "    true_maxes   = true_box_xy + true_wh_half\n",
    "    \n",
    "    pred_wh_half = pred_box_wh / 2.\n",
    "    pred_mins    = pred_box_xy - pred_wh_half\n",
    "    pred_maxes   = pred_box_xy + pred_wh_half  \n",
    "    \n",
    "    # get the corners of the intersect area\n",
    "    intersect_mins  = tf.maximum(pred_mins,  true_mins)\n",
    "    intersect_maxes = tf.minimum(pred_maxes, true_maxes)\n",
    "    intersect_wh    = tf.maximum(intersect_maxes - intersect_mins, 0.)\n",
    "    intersect_areas = intersect_wh[..., 0] * intersect_wh[..., 1]\n",
    "    \n",
    "    # get the area of the boxes\n",
    "    true_areas = true_box_wh[..., 0] * true_box_wh[..., 1]\n",
    "    pred_areas = pred_box_wh[..., 0] * pred_box_wh[..., 1]\n",
    "\n",
    "    # calculate the IOU\n",
    "    union_areas = pred_areas + true_areas - intersect_areas\n",
    "    iou_scores  = tf.truediv(intersect_areas, union_areas + 1e-16)\n",
    "    \n",
    "    # only use the IOU from boxes which actually have ROIs\n",
    "    mask = tf.equal(y_true[...,0], 1)\n",
    "    use_iou = tf.boolean_mask(1 - iou_scores, mask)\n",
    "    \n",
    "    return tf.reduce_sum(use_iou)\n",
    "    \n",
    "# create network and compiler\n",
    "model = create_network(input_size=IMAGE_SIZE, channels=24, n_blocks=1, depth=4)\n",
    "\n",
    "model.compile(optimizer='adam',\n",
    "              loss=loss_fn,\n",
    "              metrics=[binary_accuracy, overlap_iou, iou_loss_fn, adj_mse, binary_cross_entropy])\n",
    "\n",
    "# cosine learning rate annealing\n",
    "def exp_decay(x):\n",
    "    lr0 = 0.001\n",
    "    epochs_drop = 10\n",
    "    drop = 0.75\n",
    "    lrate = lr0 * math.pow(drop, math.floor((1+x)/epochs_drop))\n",
    "    return lrate\n",
    "\n",
    "learning_rate = tf.keras.callbacks.LearningRateScheduler(exp_decay)\n",
    "checkpoint = keras.callbacks.ModelCheckpoint(CHECKPOINT_PATH, monitor='val_loss', verbose=1, save_best_only=False, save_weights_only=True, mode='auto', period=1)\n",
    "\n",
    "# create train and validation generators\n",
    "folder = './stage_1_train_images'\n",
    "train_gen = generator(folder, train_filenames, pneumonia_locations, batch_size=BATCH_SIZE, image_size=IMAGE_SIZE, shuffle=True, augment=True, predict=False)\n",
    "valid_gen = generator(folder, valid_filenames, pneumonia_locations, batch_size=BATCH_SIZE, image_size=IMAGE_SIZE, shuffle=False, predict=False)\n",
    "\n",
    "print(model.summary())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "jSSSpmNE-L1L"
   },
   "outputs": [],
   "source": [
    "model.load_weights(CHECKPOINT_PATH)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 527
    },
    "colab_type": "code",
    "id": "fM5VeiuDcxUY",
    "outputId": "c3d66d6d-3c06-4aa6-f58d-46cfa42cdd9f"
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/home/eric/.local/lib/python3.5/site-packages/skimage/transform/_warps.py:110: UserWarning: Anti-aliasing will be enabled by default in skimage 0.15 to avoid aliasing artifacts when down-sampling images.\n",
      "  warn(\"Anti-aliasing will be enabled by default in skimage 0.15 to \"\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/10\n",
      "1650/1651 [============================>.] - ETA: 0s - loss: 65.6894 - binary_accuracy: 0.9027 - overlap_iou: 0.0073 - iou_loss_fn: 7.2112 - adj_mse: 5.3115 - binary_cross_entropy: 28.5878\n",
      "Epoch 00001: saving model to yolo8_2_512.h5\n",
      "1651/1651 [==============================] - 1177s 713ms/step - loss: 65.6965 - binary_accuracy: 0.9027 - overlap_iou: 0.0073 - iou_loss_fn: 7.2112 - adj_mse: 5.3112 - binary_cross_entropy: 28.5968 - val_loss: 55.5458 - val_binary_accuracy: 0.9103 - val_overlap_iou: 0.0071 - val_iou_loss_fn: 6.8071 - val_adj_mse: 4.7815 - val_binary_cross_entropy: 22.6469\n",
      "Epoch 2/10\n",
      "1650/1651 [============================>.] - ETA: 0s - loss: 59.5889 - binary_accuracy: 0.9026 - overlap_iou: 0.0075 - iou_loss_fn: 7.0337 - adj_mse: 5.0039 - binary_cross_entropy: 26.1549\n",
      "Epoch 00002: saving model to yolo8_2_512.h5\n",
      "1651/1651 [==============================] - 1179s 714ms/step - loss: 59.6172 - binary_accuracy: 0.9026 - overlap_iou: 0.0075 - iou_loss_fn: 7.0313 - adj_mse: 5.0070 - binary_cross_entropy: 26.1702 - val_loss: 56.3638 - val_binary_accuracy: 0.9446 - val_overlap_iou: 0.0076 - val_iou_loss_fn: 3.8796 - val_adj_mse: 4.6087 - val_binary_cross_entropy: 28.2971\n",
      "Epoch 3/10\n",
      "1650/1651 [============================>.] - ETA: 0s - loss: 58.3187 - binary_accuracy: 0.9090 - overlap_iou: 0.0076 - iou_loss_fn: 6.9872 - adj_mse: 4.9457 - binary_cross_entropy: 25.6911\n",
      "Epoch 00003: saving model to yolo8_2_512.h5\n",
      "1651/1651 [==============================] - 1178s 714ms/step - loss: 58.3143 - binary_accuracy: 0.9090 - overlap_iou: 0.0076 - iou_loss_fn: 6.9874 - adj_mse: 4.9458 - binary_cross_entropy: 25.6861 - val_loss: 54.3450 - val_binary_accuracy: 0.8830 - val_overlap_iou: 0.0074 - val_iou_loss_fn: 8.7051 - val_adj_mse: 4.6394 - val_binary_cross_entropy: 21.4555\n",
      "Epoch 4/10\n",
      "1650/1651 [============================>.] - ETA: 0s - loss: 57.1889 - binary_accuracy: 0.9112 - overlap_iou: 0.0076 - iou_loss_fn: 6.8760 - adj_mse: 4.9035 - binary_cross_entropy: 24.9415\n",
      "Epoch 00004: saving model to yolo8_2_512.h5\n",
      "1651/1651 [==============================] - 1179s 714ms/step - loss: 57.1921 - binary_accuracy: 0.9113 - overlap_iou: 0.0076 - iou_loss_fn: 6.8759 - adj_mse: 4.9039 - binary_cross_entropy: 24.9429 - val_loss: 51.5872 - val_binary_accuracy: 0.9285 - val_overlap_iou: 0.0075 - val_iou_loss_fn: 5.5164 - val_adj_mse: 4.5666 - val_binary_cross_entropy: 22.4251\n",
      "Epoch 5/10\n",
      "1650/1651 [============================>.] - ETA: 0s - loss: 55.2924 - binary_accuracy: 0.9098 - overlap_iou: 0.0077 - iou_loss_fn: 6.7980 - adj_mse: 4.7321 - binary_cross_entropy: 23.9342\n",
      "Epoch 00005: saving model to yolo8_2_512.h5\n",
      "1651/1651 [==============================] - 1178s 714ms/step - loss: 55.3013 - binary_accuracy: 0.9098 - overlap_iou: 0.0077 - iou_loss_fn: 6.7979 - adj_mse: 4.7320 - binary_cross_entropy: 23.9439 - val_loss: 50.9609 - val_binary_accuracy: 0.9024 - val_overlap_iou: 0.0075 - val_iou_loss_fn: 7.3245 - val_adj_mse: 4.3432 - val_binary_cross_entropy: 21.0575\n",
      "Epoch 6/10\n",
      "1650/1651 [============================>.] - ETA: 0s - loss: 53.4311 - binary_accuracy: 0.9122 - overlap_iou: 0.0080 - iou_loss_fn: 6.7136 - adj_mse: 4.4850 - binary_cross_entropy: 23.3253\n",
      "Epoch 00006: saving model to yolo8_2_512.h5\n",
      "1651/1651 [==============================] - 1180s 714ms/step - loss: 53.4355 - binary_accuracy: 0.9122 - overlap_iou: 0.0080 - iou_loss_fn: 6.7132 - adj_mse: 4.4850 - binary_cross_entropy: 23.3297 - val_loss: 47.3339 - val_binary_accuracy: 0.9019 - val_overlap_iou: 0.0084 - val_iou_loss_fn: 7.5105 - val_adj_mse: 3.9383 - val_binary_cross_entropy: 18.9889\n",
      "Epoch 7/10\n",
      "1650/1651 [============================>.] - ETA: 0s - loss: 50.3981 - binary_accuracy: 0.9157 - overlap_iou: 0.0083 - iou_loss_fn: 6.5426 - adj_mse: 4.1645 - binary_cross_entropy: 22.0859\n",
      "Epoch 00007: saving model to yolo8_2_512.h5\n",
      "1651/1651 [==============================] - 1160s 702ms/step - loss: 50.4088 - binary_accuracy: 0.9157 - overlap_iou: 0.0084 - iou_loss_fn: 6.5426 - adj_mse: 4.1661 - binary_cross_entropy: 22.0886 - val_loss: 45.0464 - val_binary_accuracy: 0.9040 - val_overlap_iou: 0.0086 - val_iou_loss_fn: 7.5902 - val_adj_mse: 3.7756 - val_binary_cross_entropy: 17.4750\n",
      "Epoch 8/10\n",
      "1650/1651 [============================>.] - ETA: 0s - loss: 47.7207 - binary_accuracy: 0.9181 - overlap_iou: 0.0087 - iou_loss_fn: 6.3768 - adj_mse: 3.8808 - binary_cross_entropy: 20.9191\n",
      "Epoch 00008: saving model to yolo8_2_512.h5\n",
      "1651/1651 [==============================] - 1159s 702ms/step - loss: 47.7336 - binary_accuracy: 0.9181 - overlap_iou: 0.0087 - iou_loss_fn: 6.3762 - adj_mse: 3.8828 - binary_cross_entropy: 20.9227 - val_loss: 40.7003 - val_binary_accuracy: 0.9381 - val_overlap_iou: 0.0090 - val_iou_loss_fn: 5.0643 - val_adj_mse: 3.3371 - val_binary_cross_entropy: 17.9522\n",
      "Epoch 9/10\n",
      "1650/1651 [============================>.] - ETA: 0s - loss: 45.8265 - binary_accuracy: 0.9219 - overlap_iou: 0.0090 - iou_loss_fn: 6.2248 - adj_mse: 3.7534 - binary_cross_entropy: 19.8807\n",
      "Epoch 00009: saving model to yolo8_2_512.h5\n",
      "1651/1651 [==============================] - 1177s 713ms/step - loss: 45.8192 - binary_accuracy: 0.9219 - overlap_iou: 0.0090 - iou_loss_fn: 6.2252 - adj_mse: 3.7532 - binary_cross_entropy: 19.8738 - val_loss: 38.3023 - val_binary_accuracy: 0.9285 - val_overlap_iou: 0.0096 - val_iou_loss_fn: 5.6310 - val_adj_mse: 3.2021 - val_binary_cross_entropy: 15.8074\n",
      "Epoch 10/10\n",
      "1650/1651 [============================>.] - ETA: 0s - loss: 43.0426 - binary_accuracy: 0.9255 - overlap_iou: 0.0094 - iou_loss_fn: 5.9794 - adj_mse: 3.5580 - binary_cross_entropy: 18.5094\n",
      "Epoch 00010: saving model to yolo8_2_512.h5\n",
      "1651/1651 [==============================] - 1178s 714ms/step - loss: 43.0400 - binary_accuracy: 0.9255 - overlap_iou: 0.0094 - iou_loss_fn: 5.9797 - adj_mse: 3.5585 - binary_cross_entropy: 18.5042 - val_loss: 36.7045 - val_binary_accuracy: 0.9293 - val_overlap_iou: 0.0097 - val_iou_loss_fn: 5.5259 - val_adj_mse: 3.1305 - val_binary_cross_entropy: 14.6822\n"
     ]
    }
   ],
   "source": [
    "history = model.fit_generator(train_gen, validation_data=valid_gen, callbacks=[learning_rate, checkpoint], epochs=10, shuffle=True, verbose=1, initial_epoch=0)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "7tyG7P79Y42U"
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/usr/local/lib/python3.5/dist-packages/tensorflow/python/ops/gradients_impl.py:100: UserWarning: Converting sparse IndexedSlices to a dense Tensor of unknown shape. This may consume a large amount of memory.\n",
      "  \"Converting sparse IndexedSlices to a dense Tensor of unknown shape. \"\n",
      "/home/eric/.local/lib/python3.5/site-packages/skimage/transform/_warps.py:110: UserWarning: Anti-aliasing will be enabled by default in skimage 0.15 to avoid aliasing artifacts when down-sampling images.\n",
      "  warn(\"Anti-aliasing will be enabled by default in skimage 0.15 to \"\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 11/15\n",
      "1650/1651 [============================>.] - ETA: 0s - loss: 30.6779 - binary_accuracy: 0.9028 - overlap_iou: 0.0101 - iou_loss_fn: 2.6471 - adj_mse: 3.4927 - binary_cross_entropy: 14.3006\n",
      "Epoch 00011: saving model to yolo8_2_512.h5\n",
      "1651/1651 [==============================] - 1179s 714ms/step - loss: 30.6764 - binary_accuracy: 0.9028 - overlap_iou: 0.0101 - iou_loss_fn: 2.6466 - adj_mse: 3.4924 - binary_cross_entropy: 14.3011 - val_loss: 26.9403 - val_binary_accuracy: 0.9025 - val_overlap_iou: 0.0105 - val_iou_loss_fn: 2.3078 - val_adj_mse: 3.1161 - val_binary_cross_entropy: 12.3747\n",
      "Epoch 12/15\n",
      " 499/1651 [========>.....................] - ETA: 12:38 - loss: 29.4109 - binary_accuracy: 0.9082 - overlap_iou: 0.0104 - iou_loss_fn: 2.5472 - adj_mse: 3.4120 - binary_cross_entropy: 13.4947"
     ]
    }
   ],
   "source": [
    "history = model.fit_generator(train_gen, validation_data=valid_gen, callbacks=[learning_rate, checkpoint], epochs=15, shuffle=True, verbose=1, initial_epoch=10)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 269
    },
    "colab_type": "code",
    "id": "y_bLpJPBcxUh",
    "outputId": "49c438f6-aabc-41b9-f98e-1bf8519da3cc"
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/home/eric/.local/lib/python3.5/site-packages/skimage/transform/_warps.py:110: UserWarning: Anti-aliasing will be enabled by default in skimage 0.15 to avoid aliasing artifacts when down-sampling images.\n",
      "  warn(\"Anti-aliasing will be enabled by default in skimage 0.15 to \"\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1440x360 with 5 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.figure(figsize=(20,5))\n",
    "plt.subplot(151)\n",
    "plt.plot(history.epoch, history.history[\"loss\"], label=\"Train loss\")\n",
    "plt.plot(history.epoch, history.history[\"val_loss\"], label=\"Valid loss\")\n",
    "plt.legend()\n",
    "plt.subplot(152)\n",
    "plt.plot(history.epoch, history.history[\"iou_loss_fn\"], label=\"Train IOU loss\")\n",
    "plt.plot(history.epoch, history.history[\"val_iou_loss_fn\"], label=\"Valid IOU loss\")\n",
    "plt.legend()\n",
    "plt.subplot(153)\n",
    "plt.plot(history.epoch, history.history[\"adj_mse\"], label=\"Train MSE loss\")\n",
    "plt.plot(history.epoch, history.history[\"val_adj_mse\"], label=\"Valid MSE loss\")\n",
    "plt.legend()\n",
    "plt.subplot(154)\n",
    "plt.plot(history.epoch, history.history[\"binary_accuracy\"], label=\"Train Confidence accuracy\")\n",
    "plt.plot(history.epoch, history.history[\"val_binary_accuracy\"], label=\"Valid Confidence accuracy\")\n",
    "plt.legend()\n",
    "plt.subplot(155)\n",
    "plt.plot(history.epoch, history.history[\"overlap_iou\"], label=\"Train iou\")\n",
    "plt.plot(history.epoch, history.history[\"val_overlap_iou\"], label=\"Valid iou\")\n",
    "plt.legend()\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "plt.figure(figsize=(20,5))\n",
    "plt.subplot(151)\n",
    "plt.plot(history.epoch, history.history[\"loss\"], label=\"Train loss\")\n",
    "plt.plot(history.epoch, history.history[\"val_loss\"], label=\"Valid loss\")\n",
    "plt.legend()\n",
    "plt.subplot(152)\n",
    "plt.plot(history.epoch, history.history[\"iou_loss_fn\"], label=\"Train IOU loss\")\n",
    "plt.plot(history.epoch, history.history[\"val_iou_loss_fn\"], label=\"Valid IOU loss\")\n",
    "plt.legend()\n",
    "plt.subplot(153)\n",
    "plt.plot(history.epoch, history.history[\"adj_mse\"], label=\"Train MSE loss\")\n",
    "plt.plot(history.epoch, history.history[\"val_adj_mse\"], label=\"Valid MSE loss\")\n",
    "plt.legend()\n",
    "plt.subplot(154)\n",
    "plt.plot(history.epoch, history.history[\"binary_accuracy\"], label=\"Train Confidence accuracy\")\n",
    "plt.plot(history.epoch, history.history[\"val_binary_accuracy\"], label=\"Valid Confidence accuracy\")\n",
    "plt.legend()\n",
    "plt.subplot(155)\n",
    "plt.plot(history.epoch, history.history[\"overlap_iou\"], label=\"Train iou\")\n",
    "plt.plot(history.epoch, history.history[\"val_overlap_iou\"], label=\"Valid iou\")\n",
    "plt.legend()\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "7dpxfbugzcLT"
   },
   "source": [
    "## Sample Predictions"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [],
   "source": [
    "def unnorm(val, idx, cell_size=256):\n",
    "    x = (val * cell_size) + (cell_size * idx)\n",
    "    \n",
    "    return x"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "c-eU6Le2zdtF"
   },
   "outputs": [],
   "source": [
    "THRESHOLD = 0.5\n",
    "OVERLAP = 0.3\n",
    "# look at some sample predictions\n",
    "samples = np.random.choice(valid_filenames, size=8, replace=False)\n",
    "\n",
    "coords = np.arange(0, 1024, 256)\n",
    "\n",
    "for filename in samples:\n",
    "    # load the image\n",
    "    img = pydicom.dcmread(os.path.join(train_dicom_dir, filename)).pixel_array\n",
    "    \n",
    "    filename = filename.split('.')[0]\n",
    "    \n",
    "    f, ax = plt.subplots(figsize=(6, 6))\n",
    "    plt.imshow(img)\n",
    "    # get boxes for the truth\n",
    "    if filename in pneumonia_locations:\n",
    "        locs = pneumonia_locations[filename]\n",
    "        for i in range(4):\n",
    "            for j in range(4):\n",
    "                pixel_data = locs[i,j,:]\n",
    "#                 print(pixel_data)\n",
    "                if pixel_data[0] > 0.5:\n",
    "                    x, y, w, h = pixel_data[1:]\n",
    "                    print(\"Truth:\", i, j)\n",
    "                    print(\"Truth:\", x, y, w, h)\n",
    "                    \n",
    "                    # unnormalize the data\n",
    "                    w = w * 1024\n",
    "                    h = h * 1024\n",
    "                    \n",
    "                    x = unnorm(x, j, 256)\n",
    "                    y = unnorm(y, i, 256)\n",
    "                    \n",
    "                    # get the corners\n",
    "                    x = x - (w // 2)\n",
    "                    y = y - (h // 2)\n",
    "                    \n",
    "                    x = int(x)\n",
    "                    y = int(y)\n",
    "                    w = int(w)\n",
    "                    h = int(h)\n",
    "                    \n",
    "                    print(\"Truth:\", x, y, w, h)\n",
    "                    \n",
    "                    rect = patches.Rectangle((x,y),w,h,linewidth=1,edgecolor='b',facecolor='none')\n",
    "                    ax.add_patch(rect)\n",
    "    \n",
    "    # predict the image\n",
    "    img = resize(img, (IMAGE_SIZE, IMAGE_SIZE), mode='reflect')\n",
    "    yhat = model.predict(img.reshape(1, IMAGE_SIZE, IMAGE_SIZE, 1))\n",
    "#     print(\"Raw:\",yhat)\n",
    "    yhat = sigmoid(yhat)\n",
    "#     print(\"Sig:\", yhat)\n",
    "    conf = np.squeeze(yhat)[:,:,0]\n",
    "    bboxes = np.squeeze(yhat)[:,:,1:]\n",
    "    boxes = []\n",
    "    \n",
    "    # loop through our predictions\n",
    "    for i in range(4):\n",
    "        for j in range(4):\n",
    "            conf_ = conf[i,j]\n",
    "#             print(\"C:\", conf_)\n",
    "            # if we have a prediction\n",
    "            if conf_ > THRESHOLD:\n",
    "                x,y,w,h = bboxes[i,j,:]\n",
    "                print(\"Pred:\", i, j, \"conf:\", conf_)\n",
    "                \n",
    "                # unnormalize the data\n",
    "                w = w * 1024\n",
    "                h = h * 1024\n",
    "\n",
    "                x = unnorm(x, j, 256)\n",
    "                y = unnorm(y, i, 256)\n",
    "                \n",
    "                # convert to upper left corner from center\n",
    "                x = np.maximum(x - (w // 2), 0)\n",
    "                y = np.maximum(y - (h // 2), 0)\n",
    "                \n",
    "                x = int(x)\n",
    "                y = int(y)\n",
    "                w = int(w)\n",
    "                h = int(h)\n",
    "                \n",
    "                # if the boxes have width and height\n",
    "                if w > 30 and h > 30:\n",
    "                    boxes.append([x,y,w,h])\n",
    "    \n",
    "    # do non-max suppression of our boxes\n",
    "    boxes = non_max_suppression_fast(np.array(boxes), OVERLAP)\n",
    "    \n",
    "    # plot our boxes\n",
    "    for box in boxes:\n",
    "        x,y,w,h = box\n",
    "        print(x, y, w, h)\n",
    "        rect = patches.Rectangle((x,y),w,h,linewidth=1,edgecolor='r',facecolor='none')\n",
    "        ax.add_patch(rect)\n",
    "    \n",
    "    for item in coords:\n",
    "        plt.axvline(item, linewidth=0.5)\n",
    "        plt.axhline(item, linewidth=0.5)\n",
    "    \n",
    "    plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "3tAntxescxUq"
   },
   "source": [
    "# Predict test images"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 34
    },
    "colab_type": "code",
    "id": "fnu7ebfScxUu",
    "outputId": "2e91d334-3204-4213-b74c-4c112b1c3512"
   },
   "outputs": [],
   "source": [
    "THRESHOLD = 0.5\n",
    "\n",
    "# load and shuffle filenames\n",
    "folder = './stage_1_test_images'\n",
    "test_filenames = os.listdir(folder)\n",
    "print('n test samples:', len(test_filenames))\n",
    "\n",
    "# create test generator with predict flag set to True\n",
    "test_gen = generator(folder, test_filenames, None, batch_size=3, image_size=IMAGE_SIZE, shuffle=False, predict=True)\n",
    "\n",
    "# create submission dictionary\n",
    "submission_dict = {}\n",
    "# loop through testset\n",
    "for imgs, filenames in test_gen:\n",
    "    \n",
    "    # predict batch of images\n",
    "    conf_preds, bbox_preds = model.predict(imgs)\n",
    "    \n",
    "    # loop through batch\n",
    "    for confs, bboxes, filename in zip(conf_preds, bbox_preds, filenames):\n",
    "        predictionString = \"\"\n",
    "        print(filename)\n",
    "        boxes = []\n",
    "        for i in range(4):\n",
    "            for j in range(4):\n",
    "                conf = confs[i, j, 0]\n",
    "                if conf > THRESHOLD:\n",
    "                    x, y, w, h = bboxes[i,j]\n",
    "                    \n",
    "                    # possible thresholds to keep our boxes within reasonable sizes?\n",
    "                    if True: #w < 600 and h < 1000:\n",
    "                        w = w * 1024\n",
    "                        h = h * 1024\n",
    "\n",
    "                        x = unnorm(x, j, 256)\n",
    "                        y = unnorm(y, i, 256)\n",
    "\n",
    "                        # convert to upper left corner from center\n",
    "                        x = x - (w // 2)\n",
    "                        y = y - (h // 2)\n",
    "                        \n",
    "                        if w > 20 and h > 20:\n",
    "                            # make sure our boxes don't run off the edges of the images\n",
    "                            w = np.minimum(w, 1024 - x)\n",
    "                            h = np.minimum(h, 1024 - y)\n",
    "                            boxes.append([x,y,w,h])\n",
    "\n",
    "        # do our non-max suppression here\n",
    "        boxes = non_max_suppression_fast(np.array(boxes), 0.3)\n",
    "        \n",
    "        # loop through our suppressed boxes and creat the prediction string\n",
    "        for box in boxes:\n",
    "            x,y,w,h = box\n",
    "            \n",
    "            x = int(x)\n",
    "            y = int(y)\n",
    "            w = int(w)\n",
    "            h = int(h)\n",
    "        \n",
    "            # create the prediction string\n",
    "            predictionString += str(conf) + ' ' + str(x) + ' ' + str(y) + ' ' + str(w) + ' ' + str(h) + ' '\n",
    "            \n",
    "        # add filename and predictionString to dictionary\n",
    "        filename = filename.split('.')[0]\n",
    "        submission_dict[filename] = predictionString\n",
    "\n",
    "    # stop if we've got them all\n",
    "    if len(submission_dict) >= len(test_filenames):\n",
    "        break\n",
    "    break\n",
    "    \n",
    "# save dictionary as csv file\n",
    "sub = pd.DataFrame.from_dict(submission_dict,orient='index')\n",
    "sub.index.names = ['patientId']\n",
    "sub.columns = ['PredictionString']\n",
    "\n",
    "now = datetime.datetime.now()\n",
    "today = str(now)[:10]\n",
    "submission_file = today + \"_yolo_submission.csv\" \n",
    "sub.to_csv(submission_file)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 34
    },
    "colab_type": "code",
    "id": "U1NEg9QCcxU4",
    "outputId": "c7e7ccf3-3e94-4996-a07d-a1098b69f3ce"
   },
   "outputs": [],
   "source": [
    "!kaggle competitions submit -c rsna-pneumonia-detection-challenge -f submission.csv -m \"YOLOv6 512x512 10 epochs\""
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 68
    },
    "colab_type": "code",
    "id": "NTQ_vDaVfvZR",
    "outputId": "a9eb81fb-572d-4236-b497-6bc44130aa89"
   },
   "outputs": [],
   "source": [
    "save_file_to_drive(\"submission.csv\", \"submission.csv\")\n",
    "save_file_to_drive(CHECKPOINT_PATH, CHECKPOINT_PATH)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 85
    },
    "colab_type": "code",
    "id": "yf06y04TVRi2",
    "outputId": "1ca39a9a-1aa6-4e2a-ec86-39a826c676dd"
   },
   "outputs": [],
   "source": [
    "# upload checkpoint to GCS\n",
    "project_id = 'mammography-198911'\n",
    "bucket_name = 'pneumonia'\n",
    "\n",
    "!gcloud config set project {project_id}\n",
    "!gsutil cp ./{CHECKPOINT_PATH} gs://{bucket_name}/"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "# upload checkpoint to GCS\n",
    "project_id = 'mammography-198911'\n",
    "bucket_name = 'pneumonia'\n",
    "\n",
    "!gcloud config set project {project_id}"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "!gsutil cp ./yolo_labels.p gs://{bucket_name}/"
   ]
  }
 ],
 "metadata": {
  "accelerator": "GPU",
  "colab": {
   "collapsed_sections": [],
   "name": "best_cnn_segmentation_connected_components_384x384_2.ipynb",
   "provenance": [],
   "version": "0.3.2"
  },
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.5.2"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 1
}