xinghaochen / awesome-hand-pose-estimation Goto Github PK

Is there a method or database with labels on visibility of annotated joints, as there are many cases that joints are annotated but not visible (enclosed by other objects). The labeling in Coco data format allows such labeling as "not labeled", "labeled but invisible", and "labeled and visible" types of annotations.

Hi i think there is a bug in MSRA evaluation code
The groundtruth file and the result of REN files has 76375 lines while there are 76391 jpg files in MSRA dataset.
Can you check?
Also, are there no additional results of other methods on MSRA dataset? Only two results are provided.

Thanks for your great job!

I need the dataset but it demands a email, and they don't accept the requirement from individual researchers or students.

Project page: https://zhengdiyu.github.io/ACR-page/
Code repo: https://github.com/ZhengdiYu/Arbitrary-Hands-3D-Reconstruction
arXiv: https://arxiv.org/abs/2303.05938

Thanks,

@xinghaochen Hello! Thank you very much for your sharing. Do you have the evaluation of RGB data set and the summary of related papers?

I find the "msra_test_list.txt" contain 76375 file, could you provide train list? What training files should I use？I found in many papers that they all said that they use theLeave one method, so they tend to train multiple models, so which model should be used?

CMU Panoptic HandDB provides 21 keypoints, not 20

handpose_x
full demo in dpcas

Hi, the results of hands'17 competition are greatly improved now:
https://competitions.codalab.org/competitions/17356#results
could you please help to find the papers in the head of the list? Thank you!

Hello!Thank you for sharing your great job! I am trying to predict with the CrossInfoNet on a kinect2 online, but I don't konw how to get these values

2019 CVPR
3D Hand Shape and Pose Estimation from a Single RGB Image
https://docs.google.com/viewer?a=v&pid=sites&srcid=ZGVmYXVsdGRvbWFpbnxnZWxpdWhhb250dXxneDo3ZjE0ZjY3OWUzYjJkYjA2

Hi Xinghao,

Thanks for adding my paper entitled "3D Hand Pose Estimation using Simulation and Partial-Supervision with a Shared Latent Space" to your list.

Just wondering if you could please update it to Oral?

Cheers.

Many of papers are using 3d images with depth, any resources using 2d ?

I think it would be helpful to mention the license, with the update of the license date , reference to the license file.

What do you think ?
I can start adding license for the rgb datasets to begin with.

A dataset called Obman Dataset is omitted
Its properties are as follows:

1. Synthetic (S)
2. RGB+ Depth
3. Obj: Interaction with objects
   4 ... ...

Hi, Thank you for your collections on hand pose estimation. I find there is your paper in the list of the arXiv papers, "Pose Guided Structured Region Ensemble Network for Cascaded Hand Pose Estimation". Can you share the related codes on the Github. Thank you again for your kindness!
regards,
weiguo

Hi , thanks for making such repo.
I have one question here:
Why do you mark "HOT-Net: Non-Autoregressive Transformer for 3D Hand-Object Pose Estimation. " as MM20 paper. I could not find the citation format -Bib Tex in Google Schoolar.
Could u explain it ? Thanks a lot.

The pdf of paper called "Weakly-supervised Domain Adaptation via GAN and Mesh Model for Estimating 3D Hand Poses Interacting Objects. " is now open source.

http://openaccess.thecvf.com/content_CVPR_2020/papers/Baek_Weakly-Supervised_Domain_Adaptation_via_GAN_and_Mesh_Model_for_Estimating_CVPR_2020_paper.pdf

Hi, Depth-Based Hand Pose Estimation: Methods, Data, and Challenges is not from IJCV2018.

Hi,

My name is Ahmed Aboukhadra from the Augmented Vision group at DFKI, I recently published a WACV paper about hand object reconstruction. Could you please add our WACV23 paper to your amazing repo on hand pose estimation? The name of the paper is THOR-Net and it's published here:

https://openaccess.thecvf.com/content/WACV2023/html/Aboukhadra_THOR-Net_End-to-End_Graformer-Based_Realistic_Two_Hands_and_Object_Reconstruction_With_WACV_2023_paper.html

There is a Github Repo for the code as well: https://github.com/ATAboukhadra/THOR-Net

Let me know if anything is missing.

Kind regards,

Ahmed

Hello,

This paper will be in the proceedings of ICCVW and it will be presented in ACVR workshop oral.
Could you please add it to the ICCV 2023 list? Below are the required links

SHOWMe: Benchmarking Object-agnostic Hand-Object 3D Reconstruction
Anilkumar Swamy, Vincent Leroy, Philippe Weinzaepfel, Fabien Baradel, Salma Galaaoui, Romain Bregier, Matthieu Armando, Jean-Sebastien Franco, Gregory Rogez

[Paper] [Project Page] [Code] [Dataset]

Thank you in Advance!

Sorry for this kind of boring question. But I cannot open the website of icvl dataset, and cannot find it anywhere.

Hello,
Can you release the code of paper SHPR-Net: Deep Semantic Hand Pose Regression From Point Clouds. I am especially interesting about the dataset preprocessed that convert from hand depth map to point cloud data. I am looking forward to your reply! Thank you very much!

Like the title suggests, I would like to propose to split the dataset section into RGB / D / RGB+D and add a column if the dataset comes with mesh annotations.

Hi Xinghao,

Thanks for maintaining this repo. It is very helpful for researchers. I want to correct two errors in your citation of my papers.

1. ECCV 2018: I typed the wrong tile on my homepage. The correct title is: "Point-to-Point Regression PointNet for 3D Hand Pose Estimation".

2. CVPR 2018: The correct title and author list are as follow:
   Hand PointNet: 3D Hand Pose Estimation using Point Sets
   Liuhao Ge, Yujun Cai, Junwu Weng, Junsong Yuan

Thank you!

see H2O

Hi, thank you for the collection of papers! This is more of suggestion than an issue. Could you also provide a comprehensive list of datasets and challenges available for hand pose estimation. Thank again for this page, it is helpful!

Great work! And thanks a lot.
Here, I want to ask if there any evaluation and comparison code for RGB hands datasets such as the code for the NYU/ICVL/MSRA datasets?

Hi, here you can find the link to the dataset that has been released for A Multi-View Video-Based 3D Hand Pose Estimation paper, that would be great if you could add it.
https://github.com/LeylaKhaleghi/MuViHand

Hi, xinghao, I have found there are a lot of paper form Hands 2017 challenge. Could you be able to download those papers? Thanks!!!

Hi,
thanks for your nice work.Do you know any dataset for two interact hands?Thanks!

Hello Xinghao,

Thanks for maintaining this useful repo. There seems to be a typo for SynHand5M dataset.
In the list of datasets, the number of joints for the SynHand5M dataset is listed as 21. However, upon further looking, I noticed they provide 1193 3D hand mesh vertices and couldn't find a mapping to extract the location of hand joints from the provided hand mesh vertices. Are you aware of the mapping between the two?

Thanks!

PeCLR: Self-Supervised 3D Hand Pose Estimation from monocular RGB viaContrastive Learning

ECCV2018
Weakly-supervised 3D Hand Pose Estimation from Monocular RGB Images
http://openaccess.thecvf.com/content_ECCV_2018/papers/Yujun_Cai_Weakly-supervised_3D_Hand_ECCV_2018_paper.pdf