'Openpose' for human pose estimation have been implemented using Tensorflow. It also provides several variants that have made some changes to the network structure for real-time processing on the CPU or low-power embedded devices.

You can even run this on your macbook with descent FPS!

Original Repo(Caffe) : https://github.com/CMU-Perceptual-Computing-Lab/openpose

CMU's Original Model on Macbook Pro 15"	Mobilenet Variant on Macbook Pro 15"	Mobilenet Variant on Jetson TK2
~0.6 FPS	~4.2 FPS @ 368x368	~10 FPS @ 368x368
2.8GHz Quad-core i7	2.8GHz Quad-core i7	Jetson TX2 Embedded Board

Implemented features are listed here : features

You need dependencies below.

• python3

• tensorflow 1.3

• opencv3

• protobuf

• python3-tk

$ git clone https://www.github.com/ildoonet/tf-openpose
$ cd tf-openpose
$ pip3 install -r requirements.txt

• cmu

  • the model based VGG pretrained network which described in the original paper.
  • I converted Weights in Caffe format to use in tensorflow.
  • weight download

• dsconv

  • Same architecture as the cmu version except for
    the depthwise separable convolution of mobilenet.
  • I trained it using 'transfer learning', but it provides not-enough speed and accuracy.

• mobilenet

  • Based on the mobilenet paper, 12 convolutional layers are used as feature-extraction layers.
  • To improve on small person, minor modification on the architecture have been made.
  • Three models were learned according to network size parameters.
    • mobilenet
      • 368x368 : weight download
    • mobilenet_fast
    • mobilenet_accurate
  • I published models which is not the best ones, but you can test them before you trained a model from the scratch.

On embedded GPU Board from Nvidia, Test results are as below.

CMU's original model can not be executed due to 'out of memory' on '368x368' size.

You can test the inference feature with a single image.

$ python3 inference.py --model=mobilenet --imgpath=...

Then you will see the screen as below with pafmap, heatmap, result and etc.

$ python3 realtime_webcam.py --camera=0 --model=mobilenet --zoom=1.0

Then you will see the realtime webcam screen with estimated poses as below. This Realtime Result was recored on macbook pro 13" with 3.1Ghz Dual-Core CPU.

See : etcs/training.md

[1] https://github.com/CMU-Perceptual-Computing-Lab/openpose

[2] Training Codes : https://github.com/ZheC/Realtime_Multi-Person_Pose_Estimation

[3] Custom Caffe by Openpose : https://github.com/CMU-Perceptual-Computing-Lab/caffe_train

[4] Keras Openpose : https://github.com/michalfaber/keras_Realtime_Multi-Person_Pose_Estimation

[1] Original Paper : https://arxiv.org/abs/1704.04861

[2] Pretrained model : https://github.com/tensorflow/models/blob/master/slim/nets/mobilenet_v1.md

[1] Tensorpack : https://github.com/ppwwyyxx/tensorpack

[1] Freeze graph : https://github.com/tensorflow/tensorflow/blob/master/tensorflow/python/tools/freeze_graph.py

[2] Optimize graph : https://codelabs.developers.google.com/codelabs/tensorflow-for-poets-2

Source: https://github.com/ildoonet/tf-pose-estimation