docs/features/image_classifier.md

Image Classifier

Overview

This module classifies whether an input image is AI-generated or a real-life photograph.

Model

• Architecture: InceptionV3
• Type: Binary Classifier (AI vs Real)
• Format: H5 model (latest-my_cnn_model.h5)

Dataset

• Total images: ~79,950
• Balanced between real and generated images
• Preprocessing: Resizing, normalization

Code Location

• Controller: features/image_classifier/controller.py
• Model Loader: features/image_classifier/model_loader.py
• Preprocessor: features/image_classifier/preprocess.py

API

• Endpoint: ENDPOINTS
• Input: Image file (PNG/JPG)
• Output: JSON response with classification result and confidence

🔙 Back to Main README