MoveNet Lightning architecture for single body pose detection.

Lightweight object detection model optimized for mobile applications.

Mobile line segment detector from NAVER Vision.

Robust video matting for alpha compositing.

CenterFace anchor-free face detection.

Sample and Computation Redistribution for Efficient Face Detection.

High performance object detector based on YOLO series.

Lightweight object classification model optimized for mobile applications.

Meet segmentation from MediaPipe.

Realtime pose detection in 3D with augmented reality.

Ultra-light fast generic face detector.

Lightweight face detection from Google MediaPipe.

Realtime single-stage face detector.

3x super resolution with lightweight CNN architecture.

Hand pose detection from MediaPipe.

Hair matte segmentation.

Single Shot Detector Lite for general object detection.

Ultra-high performance face detection architecture.

Anchor-free ultra-fast object detection.