All Projects

A real-time women safety system that uses computer vision and IoT sensors to detect threats — Fear, Assault, and Unconsciousness — through live webcam feed. The system analyzes facial expressions, eye state, and hand proximity using MediaPipe, combines it with heart rate, EEG, and motion sensor data from Arduino, and instantly triggers hardware alerts. A detailed PDF safety report is also generated for every detection.

An AI-powered web application for early detection of Lymphoma using both blood test (CBC) data and microscopic cell images. The system combines a trained XGBoost ML model for CBC analysis and a MobileNetV2 deep learning model for image classification — identifying three types of Lymphoma (CLL, FL, MCL). Results are explained by Groq's LLaMA AI, and patients can download a detailed PDF report.

A web application that identifies animal breeds instantly from a photo. Upload any animal image and the system detects the breed, shows the confidence score, and generates a detailed AI report covering the animal's origin, behavior, diet, and health. Includes a full prediction history, analytics dashboard, and a dedicated animal expert chatbot.

An IoT-powered web application for real-time monitoring of a 3-phase electrical supply using Arduino sensors. The system tracks live Voltage, Current, Power, and Transformer readings, detects faults on individual phases, manages load status, and automatically generates electricity bills with slab-based tariff breakdown — all displayed on a live dashboard with full history and CSV export.

An AI-powered real-time surveillance system that detects missing or lost children using face recognition technology. When a registered child's face is detected via webcam, the system instantly notifies parents via email and triggers a hardware alert through Arduino — all running locally with no cloud dependency.

This project, titled FakeShield, presents a complete end-to-end web-based solution for fake social-media profile detection.