Ecosyste.ms: Awesome

An open API service indexing awesome lists of open source software.

Awesome Lists | Featured Topics | Projects

https://github.com/cherubicXN/logocap

Offical code of LOGO-CAP (CVPR' 22). https://arxiv.org/abs/2109.03622
https://github.com/cherubicXN/logocap

cvpr2022 human-pose-estimation multiperson

Last synced: 2 months ago
JSON representation

Offical code of LOGO-CAP (CVPR' 22). https://arxiv.org/abs/2109.03622

Host: GitHub
URL: https://github.com/cherubicXN/logocap
Owner: cherubicXN
Created: 2022-03-26T03:31:27.000Z (almost 3 years ago)
Default Branch: main
Last Pushed: 2022-06-09T12:55:55.000Z (over 2 years ago)
Last Synced: 2024-08-04T22:13:13.379Z (6 months ago)
Topics: cvpr2022, human-pose-estimation, multiperson
Language: Python
Homepage:
Size: 2.9 MB
Stars: 33
Watchers: 4
Forks: 7
Open Issues: 1
Metadata Files:
- Readme: README.md
- Citation: citation.bib

Awesome Lists containing this project

awesome-human-pose-estimation - [code - CVPR 22, LOGO-CAP (Multi-Person 2D Pose Estimation / 2022)

README

# Learning Local-Global Contextual Adaptation for Multi-Person Pose Estimation (CVPR 2022)
**This is the offical repo of [our paper](https://arxiv.org/abs/2109.03622).**

**Abstract**: *This paper studies the problem of multi-person pose estimation in a bottom-up fashion.
With a new and strong observation that the localization issue of the center-offset formulation can be remedied in a local-window search scheme in an ideal situation, we propose a multi-person pose estimation approach, dubbed as LOGO-CAP, by learning the LOcal-GlObal Contextual Adaptation for human Pose. Specifically, our approach learns the keypoint attraction maps (KAMs) from the local keypoints expansion maps (KEMs) in small local windows in the first step, which are subsequently treated as dynamic convolutional kernels on the keypoints-focused global heatmaps for contextual adaptation, achieving accurate multi-person pose estimation. Our method is end-to-end trainable with near real-time inference speed in a single forward pass, obtaining state-of-the-art performance on the COCO keypoint benchmark for bottom-up human pose estimation. With the COCO trained model, our method also outperforms prior arts by a large margin on the challenging OCHuman dataset.*

---

## Installation (tested on Ubuntu-18.04, CUDA 11.1, pytorch-lts)

```
conda env create -f environment.yml
conda activate logocap
```

## Data Preparation

## Model Weights for Training and Testing

Download the pretrain models of HRNet backbones for training and the trained models of our LOGO-CAP with HRNet backbones by the following scripts.
```
cd weights
sh download.sh
cd ..
```

## Quickstart

After you prepared the COCO dataset on your machine, you can run the following command lines to evaluate our models on the COCO-val-2017 dataset.

```
# HRNet-W32 backbone
python tools/test.py --cfg experiments/logocap-hrnet-w32-coco.yaml --ckpt weights/logocap/logocap-hrnet-w32-coco.pth.tar

# HRNet-W48 backbone
python tools/test.py --cfg experiments/logocap-hrnet-w48-coco.yaml --ckpt weights/logocap/logocap-hrnet-w48-coco.pth.tar
```

## TODO List
- Training scripts
- README.md
- Visualization code

## Citations
If you find our work useful in your research, please consider citing:
```
@inproceedings{LOGOCAP,
title = "Learning Local-Global Contextual Adaptation for Multi-Person Pose Estimation",
author = "Nan Xue and Tianfu Wu and Gui-Song Xia and Liangpei Zhang",
booktitle = "IEEE Conference on Computer Vision and Pattern Recognition (CVPR)",
year = {2022},
}
```

## Acknowledgements

Our code is based on [DEKR](https://github.com/HRNet/DEKR). We thank Linxi Huan, Liang
Dong, Fudong Wang and the anonymous reviewers for their helpful discussions and comments.