Flask Image Classifier

Flask web app for real-time image classification using a pretrained MobileNetV2 (ImageNet weights). Upload an image, get the top class with confidence.

---

Features

• Upload any image via a Bootstrap-styled web form
• Real-time classification using MobileNetV2 pretrained on ImageNet (1000 classes)
• Predicted label + confidence percentage

Project structure

Flask-Image-Classifier/
├── app.py                # Flask application
├── templates/
│   └── index.html        # Bootstrap upload form
└── images/               # Uploaded images (auto-saved)

Quickstart

git clone https://github.com/Anjanamb/Flask-Image-Classifier.git
cd Flask-Image-Classifier
pip install flask tensorflow
python app.py

Open http://127.0.0.1:3000/, upload an image, click Predict Image.

How it works

The uploaded image is resized to 224×224, preprocessed for MobileNetV2, and decoded with decode_predictions. Example output:

Labrador retriever (95.23%)

Roadmap

• Top-K predictions (not just top-1)
• Display the uploaded image alongside the prediction
• Dockerfile
• Fine-tuning script for custom datasets

License

MIT — see anjanamb.github.io for more projects.