vnnw/person_reidentification

• Person Re-identification with OpenVINO
  • What's this
  • Reference
    • OpenVINO Toolkit and Flask Video streaming
    • OpenVINO Intel Model
  • Tested Environment
  • Required Python packages
  • How to use
  • Run app

Person Re-identification with OpenVINO

What's this

This is Person Identification Test App using Intel OpenVINO Person Re-Identification Model.

You can do followings:

• Person Detection
• Preson Re-Identification (Tracking and Counter)

Person re-identifiction - Tracking - (YouTube Link)

Person re-identifiction - Tracking and Counter- (YouTube Link)

Reference

OpenVINO Toolkit and Flask Video streaming

• Install OpenVINO Toolkit
• Flask Video streaming

OpenVINO Intel Model

• person-detection-retail-0013
• person-reidentification-retail-0031

Tested Environment

• Python 3.7.6 (need 3.6+ for f-strings)
• Windows 10 [Version 10.0.19041.388]
• OpenVINO Toolkit 2020.1+¹

Required Python packages

pip install -r requirements.txt

How to use

python app.py -h
usage: app.py [-h] -i INPUT [-d {CPU,GPU,FPGA,MYRIAD}]
              [-d_reid {CPU,GPU,FPGA,MYRIAD}] [--v4l] [-ax {0,1}] [-g GRID]

optional arguments:
  -h, --help            show this help message and exit
  -i INPUT, --input INPUT
                        Path to video file or image. 'cam' for capturing video
                        stream from camera
  -d {CPU,GPU,FPGA,MYRIAD}, --device {CPU,GPU,FPGA,MYRIAD}
                        Specify the target device for Person Detection to
                        infer on; CPU, GPU, FPGA or MYRIAD is acceptable.
  -d_reid {CPU,GPU,FPGA,MYRIAD}, --device_reidentification {CPU,GPU,FPGA,MYRIAD}
                        Specify the target device for person re-identificaiton
                        to infer on; CPU, GPU, FPGA or MYRIAD is acceptable.
  --v4l                 cv2.VideoCapture with cv2.CAP_V4L
  -ax {0,1}, --axis {0,1}
                        Specify the axis when counting person horizontally or
                        vertically (0: count horizontally(x-axis) , 1: count
                        vertically (y-axis)
  -g GRID, --grid GRID  Specify how many grid to divide frame. This is used to
                        define boundary area when the tracker counts person. 0
                        ~ 2 means not counting person. (range: 3 < max_grid)

Run app

example1. camera streaming without person counter

python app.py -i cam

example2 Specify video file with person counter

python app.py -i P1033674_1920.mp4 --grid 10 

Access the url bellow on your browser

http://127.0.0.1:5000/

Footnotes

1. openvino.inference_engine version openvino_2020.1.033 or above build does not need cpu extension.

   https://software.intel.com/en-us/forums/intel-distribution-of-openvino-toolkit/topic/848825

   ↩