skelemoa / quovadis Goto Github PK
View Code? Open in Web Editor NEWRepository for the 'Quo Vadis, Skeleton Action Recognition ?' paper
License: MIT License
Repository for the 'Quo Vadis, Skeleton Action Recognition ?' paper
License: MIT License
First at all, thanks your efforts in curating Skeletics-152 from Kinetics 700
I´m trying to draw skeletons, but it don´t know how to interpret joints relatives positions
I would appreciate if you can help me
Thanks in advance
Marcelo
I want to evaluate our method in skeletics-152, but I did not find any introduction about the configuration of 25 body joints. I would appreciate it if you could give an illustration of the configuration of 25 body joints in skeletics-152. Many thanks!
Appreciating your efforts in curating Skeleton-Mimetics from Mimetics and Skeletics-152 from Kinetics 700.
Could you please share the link to download these curated datasets (the video files and not the skeleton files)
Hi,
In the Scripts/parse_vibe_data.py, the reqd_joints are set as [38, 43, 44, 45, 46, 47, 48, 40, 37, 33,32,31, 34,35,36, 41, 39, 27,26,25, 28,29,30, 22, 19]. What are these joints? Are they are still the joints of openpose?
Best.
Can I know the frame size where joints are extracted from? I want to get relative size and position of bounding boxes .
Hi, I was wondering whether the graph structure for skeletics-152 is the same as that of NTU RGB+D. In other words, do we need to change this graph (https://github.com/kchengiva/Shift-GCN/tree/master/graph) while using skeletics-152 ?
Thanks for your releasing of this amazing work. Would you please release the SMPL pose result of VIBE for these dataset?
Hi,
I am unable to download the Skeletics 15 dataset since it would interrupt after I download only about 10 Gb. So I wonder if there is an alternative link that I can try to download the dataset.
Best
Thanks for publishing usuful datasets.
In Skeleton-Kinetics, 2d keypoint information (+ confidence) per frame is available.
In Your dataset, only 3d keypoint information (xyz by VIBE) is available?
Or 2d keypoint information(+ confidence) of openpose able to be calculated by the 3d information?
Thank you for sharing your amazing work and dataset. I have tried to download the Skeletics 152 dataset, but I always got failed-network errors during the process. I have double-checked my disk space and used different browsers, but it still always failed.
Is there any chance that you can host the file somewhere else? Any suggestions on how to download it successfully would be really helpful.
Besides, I'd like to know if the dataset also includes the original video link so that I can render the results along with the video.
Thank you so much for o your time and help.
I read that you used first 25 joints out of 49 joints. (from readme files of Skeletics and Skeletics-mimetics)
But in the given script(/Scripts/parse_vibe_data.py), you didn't used first 25 joints.
So, I want to recheck which joints did you used on the paper.
Thanks.
Thank you for sharing your amazing work and dataset. I'd like to know how you used VIBE to generate the skeleton data for Metaphorics dataset. VIBE results will become unreliable if there is no visible full-body as shown below.
Any suggestions on how you addressed this issue and generate good skeleton data even when only half-body is visible would be really appreciated. Thanks
Hi, Can you share the collected videos and annotations for the Metaphorics, please?
A declarative, efficient, and flexible JavaScript library for building user interfaces.
🖖 Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.
TypeScript is a superset of JavaScript that compiles to clean JavaScript output.
An Open Source Machine Learning Framework for Everyone
The Web framework for perfectionists with deadlines.
A PHP framework for web artisans
Bring data to life with SVG, Canvas and HTML. 📊📈🎉
JavaScript (JS) is a lightweight interpreted programming language with first-class functions.
Some thing interesting about web. New door for the world.
A server is a program made to process requests and deliver data to clients.
Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.
Some thing interesting about visualization, use data art
Some thing interesting about game, make everyone happy.
We are working to build community through open source technology. NB: members must have two-factor auth.
Open source projects and samples from Microsoft.
Google ❤️ Open Source for everyone.
Alibaba Open Source for everyone
Data-Driven Documents codes.
China tencent open source team.