← Back to Portfolio

Human Detection & Motion Analysis System

A hybrid motion recognition system that classifies human activities from video sequences using a blend of classical Computer Vision and Deep Learning. The system identifies six distinct actions: boxing, handclapping, handwaving, jogging, running, and walking.



Technologies Used


Python OpenCV 1D CNN NumPy CustomTkinter HOG & Optical Flow

Key Features

Hybrid CNN: Fuses HOG & Optical Flow for depth.

Live Prediction: 16-frame sliding window recognition.

Visual GUI: Real-time probability & motion analytics.

Smart Export: Automated logging of confidence scores.

KTH Tested