insightface

ZF/insightface

Fork 0

mirror of https://github.com/deepinsight/insightface.git synced 2026-05-16 05:27:56 +00:00

Files

History

Jia Guo 3aec9c77f4 Update README.md

2020-07-24 00:32:22 +08:00

image_infer.py

coordinateReg test code

2020-07-23 19:43:10 +08:00

README.md

Update README.md

2020-07-24 00:32:22 +08:00

README.md

Introduction

Here we provide some landmark prediction models with fast coordinate regression comparing with heatmap based methods. The input of these models is usually loose cropped face image, and the output is the direct landmark coordinates.

Pretrained Models

Model 2d106det

Given face detection bounding box, predict landmarks. Mainly used for static image inference. Please check image_infer.py for detail. Download link: TODO
Model 2d106track

Given landmarks bounding box, predict landmarks. Used for video landmarks tracking. Download link: TODO

Visualization

TODO