edvardHua/PoseEstimationForMobile

Health v2 · maintenance only

Flags: no_releases

Adoption not part of the score

1024 stars · 266 forks observed · 2026-08-28

What it is AI-extracted, prompt v1, taxonomy v1, 2026-08-30, confidence not recorded

A TensorFlow-based library implementing CPM and Hourglass models with MobileNetV2 inverted residual modules for real-time single-person human pose estimation on mobile devices. It includes training code, pretrained models, and demo apps for Android and iOS achieving roughly 60 FPS on modern hardware.

Use cases

• detect human body keypoints in real time on a phone camera feed
• build a fitness app that tracks user poses on Android or iOS
• run pose estimation models offline on mobile devices
• train a custom CPM or Hourglass pose model with TensorFlow
• benchmark lightweight pose estimation models on mobile GPUs

When to choose

• you need real-time single-person pose estimation on Android or iOS
• you want a lightweight MobileNet-based baseline for mobile inference
• you need both training code and mobile demo apps in one project

When to avoid

• you need multi-person pose estimation
• you need state-of-the-art accuracy rather than a baseline
• you target platforms other than Android or iOS
• you need actively maintained code with recent TensorFlow versions

Facets

library · maturity maintenance

machine-learning deep-learning computer-vision image-processing machine-learning computer-vision mobile-development deep-learning python cpp pose-estimation human-pose tensorflow cpm hourglass mobilenet real-time-inference tflite mace keypoint-detection android ios

Member repositories

For agents

markdown · JSON · MCP: product_card(name="edvardHua/PoseEstimationForMobile")