torch install

http://torch.ch/docs/getting-started.html

tensorflow install + GPU + Ubuntu 16.04

r0.12

sudo pip install tensorflow-gpu

export LD_LIBRARY_PATH="$LD_LIBRARY_PATH:/usr/local/cuda/lib64:/usr/local/cuda/extras/CUPTI/lib64"
export CUDA_HOME=/usr/local/cuda

$ python
Python 2.7.12 (default, Nov 19 2016, 06:48:10) 
[GCC 5.4.0 20160609] on linux2
Type "help", "copyright", "credits" or "license" for more information.
>>> import tensorflow
I tensorflow/stream_executor/dso_loader.cc:128] successfully opened CUDA library libcublas.so locally
I tensorflow/stream_executor/dso_loader.cc:128] successfully opened CUDA library libcudnn.so locally
I tensorflow/stream_executor/dso_loader.cc:128] successfully opened CUDA library libcufft.so locally
I tensorflow/stream_executor/dso_loader.cc:128] successfully opened CUDA library libcuda.so.1 locally
I tensorflow/stream_executor/dso_loader.cc:128] successfully opened CUDA library libcurand.so locally
>>> 
 

Note:

Driver mismatch

http://stackoverflow.com/questions/42288290/cuda-kernel-doesnt-match-during-installation-of-tensorflow

http://www.developersite.org/906-88688-tensorflow

opencv mask

http://stackoverflow.com/questions/7479265/apply-mask-to-image-in-opencv

inputimg.copyTo(outputimg, mask)

mask non-zero pixel will be copied from input to output

segmentation

https://arxiv.org/pdf/1411.4038.pdf

Fully Convolutional Networks for Semantic Segmentation

https://arxiv.org/abs/1611.09326v1

The One Hundred Layers Tiramisu: Fully Convolutional DenseNets for Semantic Segmentation

TF Image Segmentation: Image Segmentation framework
https://github.com/warmspringwinds/tf-image-segmentation

Human Pose Estimation

https://arxiv.org/abs/1611.00468

CRF-CNN: Modeling Structured Information in Human Pose Estimation