README.md

advanced-computer-vision-with-tensorflow

Computer vision advenced tensorflow program

Image Classification

• Multi-class Classification
  • Binary Classification (Subset of problem)
• Multi-label Classification

Recap

• Multi-class classification
• Mult-label classification

Binary vs. Multi class classification

Object Localisation

Identifying where is object in image is calling Object Localisation

Object Detection

Combined Object detection and multi-label classification is calling Object detection

Object Detection

• For each object
  • confidence score
  • bounding boxes
• Popular algorithms
  • R-CNN
  • Faster-RCNN
  • YOLO
  • SSD

Image segmentation

In digital image processing and computer vision, image segmentation is the process of partitioning a digital image into multiple image segments, also known as image regions or image objects (sets of pixels).

There are two types of segmentation:

• Semantic Segmentation: All objects are sames types from the single classification.
• Instance Segmentations:Instance Segmentation is the technique of detecting, segmenting, and classifying every individual object in an image.

Semantic vs. Instance Segmentation

Semantic Segmentation

Instance Segmentation

Useful Link

1. https://www.coursera.org/professional-certificates/tensorflow-in-practice
2. U-Net
3. DeepLab
4. Mask R-CNN

Convolutional Neural Networks Architecture

What is Transfer Learning ?

Freeze the weights

Train with learned values as default

dataset

ResNet 50

Network Architecture

Define Feature Extraction

Define Classifier

Define Input and Output