Cat-Dog Image Classifier A deep learning project to identify and differentiate cats vs dogs from images. Trained on 30k+ real images from Kaggle datasets, achieving 92% validation accuracy with CNN. Supports image uploads, webcam detection, and easy deployment.

🎯 Features Binary Classification: Cat 🐱 vs Dog 🐶 (92% accuracy) Interactive Prediction: Upload images, get instant results with confidence Live Webcam: Real-time detection (extendable) Data Augmentation: Robust to rotations, flips, zoom Transfer Learning Ready: Easy upgrade to 98%+ with MobileNetV2 Production-Ready: Streamlit/Flask deployable

📊 Performance Metric Training Validation Accuracy 91.3% 91.8% Loss 0.212 0.215 Trained 5 epochs on 24k images; extensible to 15+ epochs.

🛠️ Tech Stack Framework: TensorFlow/Keras (CNN) Data: Kaggle Cat (4GB) + Dog (2.7GB) datasets Preprocessing: ImageDataGenerator (150x150, augmentation) Hardware: Runs on CPU/GPU; ~17min/5 epochs on standard laptop

🎯 Features 92% accuracy binary classifier Interactive predictions w/ confidence Data augmentation ready Transfer learning upgrade path

Datasets Cats (4GB) Dogs (2.7GB)

🔧 Next Steps Train 15 epochs → 95%+ Add MobileNetV2 → 98%+

​