Git Product home page Git Product logo

icpr22w's Introduction

Lightweight Human Pose Estimation Using Loss Weighted by Target Heatmap

This is an official implementation of Lightweight Human Pose Estimation Using Loss Weighted by Target Heatmap that is accepted to ICPR TCAP workshop and honored as Best Paper Award (Honorable Mention). The code is developed based on the repository of MMPose.

Introduction

In this work, we lighten the computation cost and parameters of the deconvolution head network in SimpleBaseline and introduce an attention mechanism that utilizes original, inter-level, and intra-level information to intensify the accuracy. Additionally, we propose a novel loss function called heatmap weighting loss, which generates weights for each pixel on the heatmap that makes the model more focused on keypoints. Experiments demonstrate our method achieves a balance between performance, resource volume, and inference speed.

Results

results on COCO val2017

Input Size #Params GFLOPs AP AP50 AP75 APM APL AR
256 ร— 192 3.1M 0.58 65.8 87.7 74.1 62.6 72.4 72.1
384 ร— 288 3.1M 1.30 69.9 88.8 77.5 66.0 76.7 75.5

Inference Speed

Inference speed on Intel Core i7-10750H CPU and NVIDIA GTX 1650Ti (Notebooks) GPU.

Model AP FPS(GPU) FPS(CPU) GFLOPs
MobileNetV2 64.6 57.6 19.3 1.59
ShuffleNetV2 59.9 51.8 20.2 1.37
ViPNAS 67.8 22.6 4.6 0.69
Lite-HRNet18 64.8 12.8 10.3 0.20
Lite-HRNet30 67.2 7.5 6.2 0.31
Ours 65.8 55.2 18.3 0.58

speed

Installation and Preparation

Please refer to MMPose's documentation and add the file to corresponding direction.

Environment

The code is developed using python 3.7 on Ubuntu 20.04. The code is developed and tested using a single NVIDIA RTX 3090 GPU. Other platforms or GPU cards are not fully tested.

Cite us

@article{li2022lightweight,

title={Lightweight Human Pose Estimation Using Heatmap-Weighting Loss},

author={Li, Shiqi and Xiang, Xiang},

journal={arXiv preprint arXiv:2205.10611},

year={2022}

}

Acknowledgement

Thanks to:

icpr22w's People

Contributors

eglxiang avatar shi-qi-li avatar

Watchers

James Cloos avatar

Recommend Projects

  • React photo React

    A declarative, efficient, and flexible JavaScript library for building user interfaces.

  • Vue.js photo Vue.js

    ๐Ÿ–– Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.

  • Typescript photo Typescript

    TypeScript is a superset of JavaScript that compiles to clean JavaScript output.

  • TensorFlow photo TensorFlow

    An Open Source Machine Learning Framework for Everyone

  • Django photo Django

    The Web framework for perfectionists with deadlines.

  • D3 photo D3

    Bring data to life with SVG, Canvas and HTML. ๐Ÿ“Š๐Ÿ“ˆ๐ŸŽ‰

Recommend Topics

  • javascript

    JavaScript (JS) is a lightweight interpreted programming language with first-class functions.

  • web

    Some thing interesting about web. New door for the world.

  • server

    A server is a program made to process requests and deliver data to clients.

  • Machine learning

    Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.

  • Game

    Some thing interesting about game, make everyone happy.

Recommend Org

  • Facebook photo Facebook

    We are working to build community through open source technology. NB: members must have two-factor auth.

  • Microsoft photo Microsoft

    Open source projects and samples from Microsoft.

  • Google photo Google

    Google โค๏ธ Open Source for everyone.

  • D3 photo D3

    Data-Driven Documents codes.