Head-Counter

Tensorflow Object Detection API head count

事前準備

- install protoc
$ curl -OL https://github.com/google/protobuf/releases/download/v3.4.0/protoc-3.4.0-linux-x86_64.zip
$ unzip protoc-3.4.0-linux-x86_64.zip -d protoc3
$ sudo mv protoc3/bin/* /usr/local/bin/
$ sudo mv protoc3/include/* /usr/local/include/
$ chown [sudo user] /usr/local/bin/protoc
$ chown -R [sudo user] /usr/local/include/google


- install cuda9.0
$ wget https://developer.nvidia.com/compute/cuda/9.0/Prod/local_installers/cuda-repo-ubuntu1604-9-0-local_9.0.176-1_amd64-deb
$ dpkg -i cuda-repo-ubuntu1604-9-0-local_9.0.176-1_amd64-deb
$ apt-key add /var/cuda-repo-9-0-local/7fa2af80.pub
$ apt-get update
$ apt-get install cuda



- install tensorflow
$ pip3 install tensorflow-gpu==1.5

# Tensorflow Object Detection API
$ pip3 install pillow
$ pip3 install lxml
$ pip3 install jupyter
$ pip3 install matplotlib
$ git clone https://github.com/tensorflow/models
$ cd models/research
$ export PYTHONPATH=$PYTHONPATH:`pwd`:`pwd`/slim
$ export LD_LIBRARY_PATH=/usr/local/cuda-9.0/lib64${LD_LIBRARY_PATH:+:${LD_LIBRARY_PATH}}:/usr/lib/nvidia-396/
$ protoc object_detection/protos/*.proto --python_out=.

新增 main program 以及修改visualization_utils.py:

1. 把 main.py 放到 object_detection 底下
2. 修改 把visualization_utils.py 放到models/research/object_detection/utils 底下,覆蓋掉原本的
3. 執行 models/research/object_detection/utils/main.py
    $ python3 main.py  

send json

$ cd /var/www
$ vim result.php
<?php
  $filename = "models/research/object_detection/result.txt";
  $str="";
  header('Content-Type: application/json');
  if(file_exists($filename)){
  $file = fopen($filename, "r");
    if($file != NULL){
      $str .= fgets($file);
      echo json_encode($str, JSON_NUMERIC_CHECK);
    }
  }
?>

Source

• tensorflow/models

Docker版本

docker build -t "head-counter" .

docker run -it --rm head-counter bash

License

Apache License 2.0