Cats and Dogs Classifier Using CNN

Project Overview

This project implements a Convolutional Neural Network (CNN) using PyTorch to classify images of cats and dogs. The model learns visual patterns from thousands of labeled images and predicts whether a given image contains a cat or a dog. A simple GUI built with Tkinter allows users to upload an image and receive real-time predictions.

---

Model Used

The project uses a Convolutional Neural Network (CNN) implemented in PyTorch. The architecture contains:

• Convolution layers for feature extraction
• ReLU activation
• Max pooling layers
• Fully connected layers for classification

The model outputs two classes:

• Cat
• Dog

---

Time Taken

Training time on different systems:

Mac Mini M4 Pro

• ~30 minutes to train the dataset

Windows (Intel i5)

• ~2 hours training time

---

Dataset Information

Dataset Size:

~800 MB

Number of Images:

• Dogs: ~12,000 images
• Cats: ~12,000 images

Total dataset size: ~24,000 images

---

Project Directory Structure

Cats-Vs-Dogs-Classifier/
│
├── data/
│   ├── train/
│   │   ├── Cat/
│   │   └── Dog/
│
├── models/
│   └── cnn_model.py
│
├── images/
│   ├── gui_demo.png
│   ├── dog.jpg
│   └── cat.jpg
|__ clean_dataset.py
├── train.py
├── predict.py
├── predict_gui.py
├── requirements.txt
└── README.md

---

How the Model Works

1. Images are loaded from the dataset directory.
2. Each image is resized and converted into tensors.
3. The CNN extracts important visual features using convolution layers.
4. Fully connected layers classify the extracted features.
5. The model outputs the predicted label Cat or Dog.

---

Installing Virtual Environment

Mac / Linux

Create virtual environment:

python3 -m venv .venv

Activate environment:

source .venv/bin/activate

---

Windows

Create virtual environment:

python -m venv .venv

Activate environment:

.venv\Scripts\activate

---

Install Requirements

pip install -r requirements.txt

---

Downloading the Dataset

In this model , the dataset is download from Microsoft Kaggle Cats and Dogs Dataset

https://download.microsoft.com/download/3/e/1/3e1c3f21-ecdb-4869-8368-6deba77b919f/kagglecatsanddogs_5340.zip

Training the Model

Run the training script:

python train.py

After training completes, the model will be saved as:

cat_dog_cnn.pth

---

Cleaning the dataset

As the dataset contains some corrupt/broken image so I have created a script to Delete these images.

run python clean_dataset.py

Predict Without GUI

You can run prediction using the command line.

Edit the image path inside:

predict.py

Change the image path:

img = Image.open("path_to_image.jpg")

Then run:

python predict.py

---

Using GUI

Run the graphical interface:

python predict_gui.py

Steps:

1. Click Upload Image
2. Select a cat or dog image
3. Click Predict
4. The model will display the prediction

---

Requirements

Main libraries used:

torch
torchvision
pillow
numpy
tkinter

---

Author

Developed as a deep learning project using PyTorch for learning computer vision concepts and CNN-based image classification.