AI-based gait analysis using deep learning and signal processing for motion understanding.
Developed a gait analysis system using OpenPose for 2D keypoint extraction from 106 videos, processed data with NumPy and Pandas to compute kinematic features such as stride length, cadence, and joint angles. Built logistic regression and deep learning models (80K parameters) achieving 72% and 88% accuracy, respectively.
