Skip to content

Commit

Permalink
Merge pull request #2908 from open-mmlab/dev-1.x
Browse files Browse the repository at this point in the history
Dev 1.x
  • Loading branch information
Tau-J authored Jan 4, 2024
2 parents efe09cd + c9ca86a commit 509441e
Show file tree
Hide file tree
Showing 182 changed files with 27,687 additions and 502 deletions.
2 changes: 1 addition & 1 deletion CITATION.cff
Original file line number Diff line number Diff line change
@@ -1,4 +1,4 @@
cff-version: 1.2.0
cff-version: 1.3.0
message: "If you use this software, please cite it as below."
authors:
- name: "MMPose Contributors"
Expand Down
32 changes: 15 additions & 17 deletions README.md
Original file line number Diff line number Diff line change
Expand Up @@ -98,21 +98,18 @@ https://user-images.githubusercontent.com/15977946/124654387-0fd3c500-ded1-11eb-

## What's New

- We have added support for two new datasets:
- Release [RTMO](/projects/rtmo), a state-of-the-art real-time method for multi-person pose estimation.

- (CVPR 2023) [UBody](https://mmpose.readthedocs.io/zh_CN/latest/model_zoo_papers/datasets.html#ubody-cvpr-2023)
- [300W-LP](https://github.com/open-mmlab/mmpose/tree/main/configs/face_2d_keypoint/topdown_heatmap/300wlp)
![rtmo](https://github.com/open-mmlab/mmpose/assets/26127467/54d5555a-23e5-4308-89d1-f0c82a6734c2)

- Support for four new algorithms:
- Release [RTMW](/configs/wholebody_2d_keypoint/rtmpose/cocktail14/rtmw_cocktail14.md) models in various sizes ranging from RTMW-m to RTMW-x. The input sizes include `256x192` and `384x288`. This provides flexibility to select the right model for different speed and accuracy requirements.

- (ICCV 2023) [MotionBERT](https://github.com/open-mmlab/mmpose/tree/main/configs/body_3d_keypoint/motionbert)
- (ICCVW 2023) [DWPose](https://github.com/open-mmlab/mmpose/tree/main/configs/wholebody_2d_keypoint/dwpose)
- (ICLR 2023) [EDPose](https://mmpose.readthedocs.io/zh_CN/latest/model_zoo/body_2d_keypoint.html#edpose-edpose-on-coco)
- (ICLR 2022) [Uniformer](https://github.com/open-mmlab/mmpose/tree/main/projects/uniformer)
- Support inference of [PoseAnything](/projects/pose_anything). Web demo is available [here](https://openxlab.org.cn/apps/detail/orhir/Pose-Anything).

- Released the first whole-body pose estimation model, RTMW, with accuracy exceeding 70 AP on COCO-Wholebody. For details, refer to [RTMPose](/projects/rtmpose/). [Try it now!](https://openxlab.org.cn/apps/detail/mmpose/RTMPose)
- Support for two new datasets:

![rtmw](https://github.com/open-mmlab/mmpose/assets/13503330/635c4618-c459-45e8-84a5-eb68cf338d00)
- (CVPR 2023) [ExLPose](https://mmpose.readthedocs.io/en/latest/dataset_zoo/2d_body_keypoint.html#exlpose-dataset)
- (ICCV 2023) [H3WB](/docs/en/dataset_zoo/3d_wholebody_keypoint.md)

- Welcome to use the [*MMPose project*](/projects/README.md). Here, you can discover the latest features and algorithms in MMPose and quickly share your ideas and code implementations with the community. Adding new features to MMPose has become smoother:

Expand All @@ -121,6 +118,8 @@ https://user-images.githubusercontent.com/15977946/124654387-0fd3c500-ded1-11eb-
- Utilize the powerful capabilities of MMPose in the form of independent projects without being constrained by the code framework.
- Newly added projects include:
- [RTMPose](/projects/rtmpose/)
- [RTMO](/projects/rtmo/)
- [PoseAnything](/projects/pose_anything/)
- [YOLOX-Pose](/projects/yolox_pose/)
- [MMPose4AIGC](/projects/mmpose4aigc/)
- [Simple Keypoints](/projects/skps/)
Expand All @@ -130,15 +129,14 @@ https://user-images.githubusercontent.com/15977946/124654387-0fd3c500-ded1-11eb-

<br/>

- October 12, 2023: MMPose [v1.2.0](https://github.com/open-mmlab/mmpose/releases/tag/v1.2.0) has been officially released, with major updates including:
- January 4, 2024: MMPose [v1.3.0](https://github.com/open-mmlab/mmpose/releases/tag/v1.3.0) has been officially released, with major updates including:

- Support for new datasets: UBody, 300W-LP.
- Support for new algorithms: MotionBERT, DWPose, EDPose, Uniformer.
- Migration of Associate Embedding, InterNet, YOLOX-Pose algorithms.
- Migration of the DeepFashion2 dataset.
- Support for Badcase visualization analysis, multi-dataset evaluation, and keypoint visibility prediction features.
- Support for new datasets: ExLPose, H3WB
- Release of new RTMPose series models: RTMO, RTMW
- Support for new algorithm PoseAnything
- Enhanced Inferencer with optional progress bar and improved affinity for one-stage methods

Please check the complete [release notes](https://github.com/open-mmlab/mmpose/releases/tag/v1.2.0) for more details on the updates brought by MMPose v1.2.0!
Please check the complete [release notes](https://github.com/open-mmlab/mmpose/releases/tag/v1.3.0) for more details on the updates brought by MMPose v1.3.0!

## 0.x / 1.x Migration

Expand Down
32 changes: 15 additions & 17 deletions README_CN.md
Original file line number Diff line number Diff line change
Expand Up @@ -96,21 +96,18 @@ https://user-images.githubusercontent.com/15977946/124654387-0fd3c500-ded1-11eb-

## 最新进展

- 我们支持了两个新的数据集:
- 发布了单阶段实时多人姿态估计模型 [RTMO](/projects/rtmo)。相比 RTMPose 在多人场景下性能更优

- (CVPR 2023) [UBody](https://mmpose.readthedocs.io/zh_CN/latest/model_zoo_papers/datasets.html#ubody-cvpr-2023)
- [300W-LP](https://github.com/open-mmlab/mmpose/tree/main/configs/face_2d_keypoint/topdown_heatmap/300wlp)
![rtmo](https://github.com/open-mmlab/mmpose/assets/26127467/54d5555a-23e5-4308-89d1-f0c82a6734c2)

- 支持四个新算法:
- 发布了不同尺寸的 [RTMW](/configs/wholebody_2d_keypoint/rtmpose/cocktail14/rtmw_cocktail14.md) 模型,满足不同的使用场景。模型尺寸覆盖从 RTMW-m 到 RTMW-x 的模型,输入图像尺寸包含 256x192 和 384x288

- (ICCV 2023) [MotionBERT](https://github.com/open-mmlab/mmpose/tree/main/configs/body_3d_keypoint/motionbert)
- (ICCVW 2023) [DWPose](https://github.com/open-mmlab/mmpose/tree/main/configs/wholebody_2d_keypoint/dwpose)
- (ICLR 2023) [EDPose](https://mmpose.readthedocs.io/zh_CN/latest/model_zoo/body_2d_keypoint.html#edpose-edpose-on-coco)
- (ICLR 2022) [Uniformer](https://github.com/open-mmlab/mmpose/tree/main/projects/uniformer)
- 支持了 [PoseAnything](/projects/pose_anything) 的推理。[在线试玩](https://openxlab.org.cn/apps/detail/orhir/Pose-Anything)

- 发布首个在 COCO-Wholebody 上精度超过 70 AP 的全身姿态估计模型 RTMW,具体请参考 [RTMPose](/projects/rtmpose/)[在线试玩](https://openxlab.org.cn/apps/detail/mmpose/RTMPose)
- 我们支持了两个新的数据集:

![rtmw](https://github.com/open-mmlab/mmpose/assets/13503330/635c4618-c459-45e8-84a5-eb68cf338d00)
- (CVPR 2023) [ExLPose](https://mmpose.readthedocs.io/en/latest/dataset_zoo/2d_body_keypoint.html#exlpose-dataset)
- (ICCV 2023) [H3WB](/docs/en/dataset_zoo/3d_wholebody_keypoint.md)

- 欢迎使用 [*MMPose 项目*](/projects/README.md)。在这里,您可以发现 MMPose 中的最新功能和算法,并且可以通过最快的方式与社区分享自己的创意和代码实现。向 MMPose 中添加新功能从此变得简单丝滑:

Expand All @@ -119,6 +116,8 @@ https://user-images.githubusercontent.com/15977946/124654387-0fd3c500-ded1-11eb-
- 通过独立项目的形式,利用 MMPose 的强大功能,同时不被代码框架所束缚
- 最新添加的项目包括:
- [RTMPose](/projects/rtmpose/)
- [RTMO](/projects/rtmo/)
- [PoseAnything](/projects/pose_anything/)
- [YOLOX-Pose](/projects/yolox_pose/)
- [MMPose4AIGC](/projects/mmpose4aigc/)
- [Simple Keypoints](/projects/skps/)
Expand All @@ -128,15 +127,14 @@ https://user-images.githubusercontent.com/15977946/124654387-0fd3c500-ded1-11eb-

<br/>

- 2023-10-12:MMPose [v1.2.0](https://github.com/open-mmlab/mmpose/releases/tag/v1.2.0) 正式发布了,主要更新包括:
- 2024-01-04:MMPose [v1.3.0](https://github.com/open-mmlab/mmpose/releases/tag/v1.3.0) 正式发布了,主要更新包括:

- 支持新数据集:UBody、300W-LP。
- 支持新算法:MotionBERT、DWPose、EDPose、Uniformer
- 迁移 Associate Embedding、InterNet、YOLOX-Pose 算法。
- 迁移 DeepFashion2 数据集。
- 支持 Badcase 可视化分析、多数据集评测、关键点可见性预测功能。
- 支持新数据集:ExLPose、H3WB
- 发布 RTMPose 系列新模型:RTMO、RTMW
- 支持新算法 PoseAnything
- 推理器 Inferencer 支持可选的进度条、提升与单阶段模型的适配性

请查看完整的 [版本说明](https://github.com/open-mmlab/mmpose/releases/tag/v1.2.0) 以了解更多 MMPose v1.2.0 带来的更新!
请查看完整的 [版本说明](https://github.com/open-mmlab/mmpose/releases/tag/v1.3.0) 以了解更多 MMPose v1.3.0 带来的更新!

## 0.x / 1.x 迁移

Expand Down
125 changes: 125 additions & 0 deletions configs/_base_/datasets/exlpose.py
Original file line number Diff line number Diff line change
@@ -0,0 +1,125 @@
dataset_info = dict(
dataset_name='exlpose',
paper_info=dict(
author='Sohyun Lee, Jaesung Rim, Boseung Jeong, Geonu Kim,'
'ByungJu Woo, Haechan Lee, Sunghyun Cho, Suha Kwak',
title='Human Pose Estimation in Extremely Low-Light Conditions',
container='arXiv',
year='2023',
homepage='https://arxiv.org/abs/2303.15410',
),
keypoint_info={
0:
dict(
name='left_shoulder',
id=0,
color=[0, 255, 0],
type='upper',
swap='right_shoulder'),
1:
dict(
name='right_shoulder',
id=1,
color=[255, 128, 0],
type='upper',
swap='left_shoulder'),
2:
dict(
name='left_elbow',
id=2,
color=[0, 255, 0],
type='upper',
swap='right_elbow'),
3:
dict(
name='right_elbow',
id=3,
color=[255, 128, 0],
type='upper',
swap='left_elbow'),
4:
dict(
name='left_wrist',
id=4,
color=[0, 255, 0],
type='upper',
swap='right_wrist'),
5:
dict(
name='right_wrist',
id=5,
color=[255, 128, 0],
type='upper',
swap='left_wrist'),
6:
dict(
name='left_hip',
id=6,
color=[0, 255, 0],
type='lower',
swap='right_hip'),
7:
dict(
name='right_hip',
id=7,
color=[255, 128, 0],
type='lower',
swap='left_hip'),
8:
dict(
name='left_knee',
id=8,
color=[0, 255, 0],
type='lower',
swap='right_knee'),
9:
dict(
name='right_knee',
id=9,
color=[255, 128, 0],
type='lower',
swap='left_knee'),
10:
dict(
name='left_ankle',
id=10,
color=[0, 255, 0],
type='lower',
swap='right_ankle'),
11:
dict(
name='right_ankle',
id=11,
color=[255, 128, 0],
type='lower',
swap='left_ankle'),
12:
dict(name='head', id=12, color=[51, 153, 255], type='upper', swap=''),
13:
dict(name='neck', id=13, color=[51, 153, 255], type='upper', swap='')
},
skeleton_info={
0: dict(link=('head', 'neck'), id=0, color=[51, 153, 255]),
1: dict(link=('neck', 'left_shoulder'), id=1, color=[51, 153, 255]),
2: dict(link=('neck', 'right_shoulder'), id=2, color=[51, 153, 255]),
3: dict(link=('left_shoulder', 'left_elbow'), id=3, color=[0, 255, 0]),
4: dict(link=('left_elbow', 'left_wrist'), id=4, color=[0, 255, 0]),
5: dict(
link=('right_shoulder', 'right_elbow'), id=5, color=[255, 128, 0]),
6:
dict(link=('right_elbow', 'right_wrist'), id=6, color=[255, 128, 0]),
7: dict(link=('neck', 'right_hip'), id=7, color=[51, 153, 255]),
8: dict(link=('neck', 'left_hip'), id=8, color=[51, 153, 255]),
9: dict(link=('right_hip', 'right_knee'), id=9, color=[255, 128, 0]),
10:
dict(link=('right_knee', 'right_ankle'), id=10, color=[255, 128, 0]),
11: dict(link=('left_hip', 'left_knee'), id=11, color=[0, 255, 0]),
12: dict(link=('left_knee', 'left_ankle'), id=12, color=[0, 255, 0]),
},
joint_weights=[
0.2, 0.2, 0.2, 1.3, 1.5, 0.2, 1.3, 1.5, 0.2, 0.2, 0.5, 0.2, 0.2, 0.5
],
sigmas=[
0.079, 0.079, 0.072, 0.072, 0.062, 0.062, 0.107, 0.107, 0.087, 0.087,
0.089, 0.089, 0.079, 0.079
])
Loading

0 comments on commit 509441e

Please sign in to comment.