Compute 2D human pose and angles from a video or a webcam.

SafeCity is an intelligent web-based safety monitoring application designed to detect real-time distress through audio, visual, motion, and contextual signals. Using modern web technologies and AI inference models, the system provides tiered safety alerts that escalate automatically when signs of distress are detected.

Home of the GaitKeeper Run Analysis Tool

Quickly add MediaPipe Pose Estimation and Detection to your iOS app. Enable powerful features in your app powered by the body or hand.

Yoga pose classification model based on pose estimation (BlazePose 3D)

🚀一个基于机器视觉 BlazePose 神经网络的Ai健身教练

Blaze Pose and YOLOv8 powered robot for real-time person tracking and fall-risk assessment. An Arduino-controlled mobile platform provides assistance to the tracked individual.

A telerehabilitation system for the OAK-D AI camera with autonomous and real-time performance feedback.

This is a real-time pose estimation project that detects 33 human body landmarks in images, videos, and live webcam streams. Built using MediaPipe, OpenCV, and Streamlit, this project provides an interactive and efficient way to analyze human movements using Blaze Pose detection method.

AI Gym Trainer

[JS/TensorFlow] JavaScript library that implements machine learning-based models for human pose estimation and human movement analysis. It allows you to easily implement three neural network models for human pose estimation: MoveNet, PoseNet, and BlazePose. No Node.js is required. TensorFlowJS was used, and an example app is included.

Compare SVM mode yoga movement classification accuracy with Linear kernel, Polynomial kernel, RBF (Radial Basis Function) kernel, LSTM with accuracy up to 98%. In addition, it also supports adjusting the practitioner's movements according to standard movements.

Realtime pose landmark detection with BlazePoseBarracuda in Unity

Automatic 3D Character animation using Pose Estimation and Landmark Generation techniques

predicting of 3d landmarks using Blazepose and Detectrin2 model base on 2d video by Python

Full body pose estimation to be used with HMD (Quest2) built in Unity

BlazePose: Body Segmentation for TFJS and NodeJS

This is the project repository for the research study Human Action Recognition using BlazePose Skeleton on Spatial Temporal Graph Convolutional Networks presented by Motasem S. Alsawadi and Miguel Rio.

A collection of notebooks for my final year project. The notebooks are used to create a virtual personal trainer to check bicep curls, squats and overhead presses.