3x super resolution with lightweight CNN architecture.

Ultra-light fast generic face detector.

Meet segmentation from MediaPipe.

Single Shot Detector Lite for general object detection.

Realtime single-stage face detector.

CenterFace anchor-free face detection.

MoveNet Lightning architecture for single body pose detection.

High performance object detector based on YOLO series.

Lightweight object classification model optimized for mobile applications.

Hand pose detection from MediaPipe.

Ultra-high performance face detection architecture.

Hair matte segmentation.

Lightweight object detection model optimized for mobile applications.

Sample and Computation Redistribution for Efficient Face Detection.

Realtime pose detection in 3D with augmented reality.

Robust video matting for alpha compositing.

Lightweight face detection from Google MediaPipe.

Anchor-free ultra-fast object detection.

Mobile line segment detector from NAVER Vision.