This repository currently implemented the Convolutional Pose Machine (CPM) using TensorFlow. Instead of normal convolution, inverted residuals (also known as Mobilenet V2) module has been used inside the CPM for faster inference. More experimental models will release as time goes by.

Hence, the respository contains:

• Code of training model
• Code of converting model to TensorFlow Lite
• Android Demo
• IOS Demo (TODO)

Below GIF is catch on Mi Mix2s (5 FPS)

Download the apk of demo.

You can buy me a coke if you think my work is helpful for you.
ETH address: 0x8fcF32D797968B64428ab2d8d09ce2f74143398E

• Python3
• TensorFlow >= 1.4

Training dataset available through google driver.

Unzip it will obtain the following file structure

# root @ ubuntu in ~/hdd/ai_challenger $ tree -L 1 . . ├── ai_challenger_train.json ├── ai_challenger_valid.json ├── train └── valid

The traing dataset only contains single person images and it come from the competition of AI Challenger. I transfer the annotation into COCO format for using the data augument code from tf-pose-estimation respository.

In training procedure, we use cfg file on experiments folder for passing the hyper-parameter.

Below is the content of mv2_cpm.cfg.

[Train] model: 'mv2_cpm' checkpoint: False datapath: '/root/hdd/ai_challenger' imgpath: '/root/hdd/' visible_devices: '0, 1, 2' multiprocessing_num: 8 max_epoch: 1000 lr: '0.001' batchsize: 5 decay_rate: 0.95 input_width: 224 input_height: 224 n_kpoints: 14 scale: 2 modelpath: '/root/hdd/trained/mv2_cpm/models' logpath: '/root/hdd/trained/mv2_cpm/log' num_train_samples: 20000 per_update_tensorboard_step: 500 per_saved_model_step: 2000 pred_image_on_tensorboard: True

The cfg not cover all the parameters of the model, there still have some parameters in the network_mv2_cpm.py.

Build the docker by the following command:

cd training/docker docker build -t single-pose .

or

docker pull edvardhua/single-pose 

Then run the following command to train the model:

nvidia-docker run -it -d \ -v <dataset_path>:/data5 -v <training_code_path>/training:/workspace \ -p 6006:6006 -e LOG_PATH=/root/hdd/trained/mv2_cpm/log \ -e PARAMETERS_FILE=experiments/mv2_cpm.cfg edvardhua/single-pose

Also, it will create the tensorboard on port 6006. Beside, make sure you install the nvidia-docker.

1. install the dependencies.

cd training pip3 install -r requirements.txt

Beside, you also need to install cocoapi

2. Edit the parameters files in experiments folder, it contains almost all the hyper-parameters and other configuration you need to define in training. After that, passing the parameters file to start the training:

cd training python3 src/train.py experiments/mv2_cpm.cfg

After 12 hour training, the model is almost coverage on 3 Nvidia 1080Ti graphics cards, below is the corresponding plot on tensorboard.

Can be download here.

After you training the model, the following command can transfer the model into tflite.

# Convert to frozen pb. cd training python3 src/gen_frozen_pb.py \ --checkpoint=<you_training_model_path>/model-xxx --output_graph=<you_output_model_path>/model-xxx.pb \ --size=224 --model=mv2_cpm_2 # Convert to tflite. # See https://github.com/tensorflow/tensorflow/blob/master/tensorflow/docs_src/mobile/tflite/devguide.md for more information. bazel-bin/tensorflow/contrib/lite/toco/toco \ --input_file=<you_output_model_path>/model-xxx.pb \ --output_file=<you_output_tflite_model_path>/mv2-cpm.tflite \ --input_format=TENSORFLOW_GRAPHDEF --output_format=TFLITE \ --inference_type=FLOAT \ --input_shape="1,224,224,3" \ --input_array='image' \ --output_array='Convolutional_Pose_Machine/stage_5_out'

Then, place the tflite file in android_demo/app/src/main/assets and modify the parameters in ImageClassifierFloatInception.kt.

...... ...... // parameters need to modify in ImageClassifierFloatInception.kt /**  * Create ImageClassifierFloatInception instance  *  * @param imageSizeX Get the image size along the x axis.  * @param imageSizeY Get the image size along the y axis.  * @param outputW The output width of model  * @param outputH The output height of model  * @param modelPath Get the name of the model file stored in Assets.  * @param numBytesPerChannel Get the number of bytes that is used to store a single  * color channel value.  */ fun create( activity: Activity,      imageSizeX: Int = 224,      imageSizeY: Int = 224,      outputW: Int = 112,      outputH: Int = 112, modelPath: String = "mv2-cpm.tflite", numBytesPerChannel: Int = 4 ): ImageClassifierFloatInception = ImageClassifierFloatInception( activity, imageSizeX, imageSizeY, outputW, outputH, modelPath, numBytesPerChannel) ...... ......

Finally, import the project to Android Studio and run in you smartphone.

If you are an IOS enthusiast who are interested in this project and want to migrate to ios, we welcome to submit a pull request.

[1] Paper of Convolutional Pose Machines
[2] Paper of MobileNet V2
[3] Repository of tf-pose-estimation
[4] Devlope guide of TensorFlow Lite

Apache License 2.0