This project implements Image classification system using Quantum Machine Learning . It combines classical image processing, quantum feature extraction using quantum circuits, and machine learning classification to categorize images from a dataset of aerial or satellite images.
🖼️ Classical Image Preprocessing • Images are converted to grayscale, resized (e.g., 32x32 pixels), normalized, and flattened into feature vectors.
⚛️ Quantum Feature Extraction • Parameterized quantum circuits encode classical image data into quantum states.
• Quantum gates (rotations and entanglement) capture high-dimensional correlations.
• Measurements generate quantum features that are difficult to model classically.
🔗 Hybrid Feature Vector • Classical and quantum features are concatenated into a hybrid vector. • This vector represents each image in a combined classical-quantum space.
🌲 Machine Learning Model • A Random Forest classifier is trained on the hybrid vectors. • The model predicts land use classes such as agricultural, beach, forest, etc. • Trained models are persisted for fast inference.
🧪 Backend API (Flask) • Endpoint: /classify • Accepts image uploads, processes them, and returns: • Predicted class • Confidence score (in JSON)!
💻 Frontend • A simple web interface for: • Uploading images • Viewing prediction results • Error handling
🚀 Significance
This project showcases a practical integration of quantum computing and classical AI in real-world tasks. Quantum-enhanced feature extraction may provide richer data representations, leading to improved classification performance.