Sphere-Meshes for Real-Time Hand Modeling and Tracking

• Video: http://lgg.epfl.ch/publications/2016/HModel/video.mp4
• Paper PDF: http://lgg.epfl.ch/publications/2016/HModel/paper.pdf
• Dataset: http://lgg.epfl.ch/publications/2016/HModel/teaser.zip

Disclaimer

To obtain the results shown in the video proper hardware is necessary:

• Windows 8
• Intel Core i7 @4GhZ
• CUDA Graphic card (NVIDIA GTX980 used in our demo)
• Inter RealSense depth camera

Other notes:

• note the software must be compiled in 64bits
• Wristband color calibration (make sure the wristband is detected robustly otherwise the tracking might not perform as effectively, you can check this by enabling "show wband" in the hmodel_atb application)

BibTex

@article{hmodel,
title = {Sphere-Meshes for Real-Time Hand Modeling and Tracking}, 
author = {Anastasia Tkach and Mark Pauly and Andrea Tagliasacchi}, 
journal = {ACM Transactions on Graphics (Proceedings of SIGGRAPH Asia)}, 
year = {2016}}

Running "Teaser" Sequence

in hmodel/apps/hmodel_atb/main.cpp

• change 21 | std::string sequence_path to the path to teaser dataset in you machine
• change 22 | std::string data_path to the path to hmodel/data/ folder on your machine
• set 17 | bool benchmarkto true

Running Live

in hmodel/apps/hmodel_atb/main.cpp

• set 17 | bool benchmark to false
• wear a blue wristband and make sure that the wristband is always in the view of the sensor
• c++ implementation of automatic hand model calibration is coming
• for now to approximately adjust the model for your hand
  • press key 1 for uniform scaling up
  • press key 2 for uniform scaling down
  • press key 3 for width scaling up
  • press key 4 for width scaling down
  • press key 5 for thickness scaling up
  • press key 6 for thickness scaling down