Orange Image Classification API & Web App

Lightweight Flask application that serves a trained convolutional neural network (TensorFlow / Keras .h5 + optimized .tflite) for orange image classification. Users can upload a photo via a clean HTML form, receive the predicted class and confidence, and run the service locally or in containers (Docker + optional Compose). Designed for fast inference, simple deployment, and easy model replacement.

Technologies Used

Key Features

• Image upload & prediction UI
• REST prediction endpoint (JSON response)
• Dual model format (.h5 & .tflite)
• Confidence score output
• Basic input validation & secure temp storage
• Containerized deployment (Dockerfile + Compose)
• Easily swappable model file