Image Classification Using Deep Learning

Overview

This project classifies images of cats and dogs using a Convolutional Neural Network (CNN) inspired by the VGG architecture.
The goal is to enable automatic identification of household animals through deep learning techniques.

---

Model Highlights

• Architecture: VGG-style CNN with stacked 3×3 convolutional filters
• Activations: tanh with Dropout regularization
• Optimizer: RMSprop
• Loss Function: Binary Cross-Entropy
• Data Augmentation: rotation, zoom, shear, horizontal flips
• Performance: Achieved ~87% validation accuracy

---

Dataset

The dataset used in this project is the Dogs vs. Cats dataset from Kaggle.
It contains a total of 10,000 images — 5,000 cats and 5,000 dogs — of varying sizes and qualities.

All images were resized to 128×128 pixels for consistency before training.
This dataset provides a balanced and widely used benchmark for binary image classification tasks.

Team

Yathartha Aarush

Smrithi Shenoy