examples/deep_hand_pose/train.sh should run my implementation of [1] on the NYU dataset. It can be configured to use a variety of datasets.

[1] M. Oberweger, P. Wohlhart, and V. Lepetit. Hands Deep in Deep Learning for Hand Pose Estimation. CVWW, 2015.

I've added 3 new layers to caffe, in the src/caffe/layers directory

• HandData which loads a variety of hand image/annotation formats
• PCA which implements Oberweger's PCA bottleneck initialization
• MVRegLoss which adds visualization to the euclidean loss

• Download the NYU HANDS dataset (http://horatio.cs.nyu.edu/mit/tompson/nyu_hand_dataset_v2.zip) into /mnt/data/NYU-Hands-v2/
• Download the deep_hand_pose code

git clone [email protected]:jsupancic/deep_hand_pose.git ~/deep_hand_pose

• I've converted the annotations from .mat to .csv for you:

cd ~/deep_hand_pose/

cp nyu_csv_annotations/test/*.csv /mnt/data/NYU-Hands-v2/test/

cp nyu_csv_annotations/test/*.csv /mnt/data/NYU-Hands-v2/train/

• Compile caffe and deep_hand_pose

cd ~/deep_hand_pose/ && mkdir build && pushd build && cmake .. && make -j16 && popd

• create the directory where results will be stored

mkdir out

• now run the pre-trained model on the NYU dataset!

examples/deep_hand_pose/train.sh

The license is the same as caffe from which this is derived. Please see below.

If you find this useful, relevant citations would be

@article{supancic2015depth,
  title={Depth-based hand pose estimation: methods, data, and challenges},
  author={Supancic III, James Steven and Rogez, Gregory and Yang, Yi and Shotton, Jamie and Ramanan, Deva},
  journal={arXiv preprint arXiv:1504.06378},
  year={2015}
}

and

@article{oberweger2015hands,
  title={Hands Deep in Deep Learning for Hand Pose Estimation},
  author={Oberweger, Markus and Wohlhart, Paul and Lepetit, Vincent},
  journal={arXiv preprint arXiv:1502.06807},
  year={2015}
}

Caffe is a deep learning framework made with expression, speed, and modularity in mind. It is developed by the Berkeley Vision and Learning Center (BVLC) and community contributors.

Check out the project site for all the details like

• DIY Deep Learning for Vision with Caffe
• Tutorial Documentation
• BVLC reference models and the community model zoo
• Installation instructions

and step-by-step examples.

Please join the caffe-users group or gitter chat to ask questions and talk about methods and models. Framework development discussions and thorough bug reports are collected on Issues.

Happy brewing!

Caffe is released under the BSD 2-Clause license. The BVLC reference models are released for unrestricted use.

Please cite Caffe in your publications if it helps your research:

@article{jia2014caffe,
  Author = {Jia, Yangqing and Shelhamer, Evan and Donahue, Jeff and Karayev, Sergey and Long, Jonathan and Girshick, Ross and Guadarrama, Sergio and Darrell, Trevor},
  Journal = {arXiv preprint arXiv:1408.5093},
  Title = {Caffe: Convolutional Architecture for Fast Feature Embedding},
  Year = {2014}
}