Skip to content

Latest commit

 

History

History
 
 

keypointdetection

Folders and files

NameName
Last commit message
Last commit date

parent directory

..
 
 
 
 
 
 
 
 

English | 简体中文

Keypoint Detection Model

Currently FastDeploy supports the deployment of two key point detection methods.

Task Description Model Format Example Version
Single person key point detection Deploy PP-TinyPose series models with input images containing only a single person Paddle Please refer to tinypose Release/2.5
Single/multi person key point detection Deploy the model tandem task of PicoDet + PP-TinyPose, where the input image first passes through the detection model to get an independent portrait sub-image, and then passes through the PP-TinyPose model to detect key points Paddle Please refer to det_keypoint_unite Release/2.5

Prepare Pre-trained Model

Here we provides the following pre-trained models, you can download and use them directly.

Model Description Model Format Version
PP-TinyPose-128x96 Single person key point detection Paddle Release/2.5
PP-TinyPose-256x192 Single person key point detection Paddle Release/2.5
PicoDet-S-Lcnet-Pedestrian-192x192 + PP-TinyPose-128x96 Single person key point detection tandem configuration Paddle Release/2.5
PicoDet-S-Lcnet-Pedestrian-320x320 + PP-TinyPose-256x192 Multi person key point detection tandem configuration Paddle Release/2.5