mercurialjd / concentration-detection

This repo takes advantages of OpenCV and TensorFlow to achieve the goal of concentration analysis (camera/video part).

Concentration Analysis (Camera/Video Part)

This repo takes advantages of OpenCV and TensorFlow to achieve the goal of concentration analysis (camera/video part).

If you prefer Chinese README, please refer to this file.

Note: The original video before being processed is referred from the Internet

Tested Prerequisites

• opencv-python 4.2.0.32
• tensorflow 1.14.0
• Python 3.6.9
• Codes from this repo has been tested on Ubuntu 18.04

Installation

git clone https://github.com/MercurialJD/Concentration-Detection.git

User's Guidance

This project accepts video files or real-time camera records. In the case no argument is provided or the given argument points to a camera, the corresponding camera will be automatically invoked; If a video file is given as an argument, the video will be analyzed.

Video Mode

For video formats that can be identified by OpenCV:

python3 concentration_estimation.py --video /path/to/your/video.mp4

Camera Mode

Default camera:

python3 concentration_estimation.py

Explicitly designated camera:

python3 concentration_estimation.py --cam 0

Output Adjustments

Please adjust concentration_estimation.py w.r.t. your requirements.

• If the processed video is needed, please uncomment following lines:

  # fourcc = cv2.VideoWriter_fourcc(*'mp4v')
  # out = cv2.VideoWriter('output.mp4', fourcc, 20.0, (width, height))
  ...
  # cv2.putText(frame, "FACING", (height//50, width//50), cv2.FONT_HERSHEY_DUPLEX, 1, (0, 0, 0), 1)
  ...
  # cv2.putText(frame, "UNFACED", (height//50, width//50), cv2.FONT_HERSHEY_DUPLEX, 1, (0, 0, 0), 1)
  ...
  # out.write(frame)
  ...
  # out.release()
  

• If real-tile preview is needed, please uncomment the following line:

  # cv2.imshow("Preview", frame)
  

• If the facial identification box need to be customized, please modify:

  pose_estimator.draw_annotation_box(...)
  

• If head-pose axes need to be customized, please modify:

  pose_estimator.draw_axis(...)
  

Basic Procedure

1. Facial recognition with DNN module provided by OpenCV
2. Facial landmarks with 68-point estimation provided by TensorFlow
3. Pose estimation with PnP algorithm
4. Stabilization with Kalman filter
5. Concentration analysis

License

Please refer to LICENSE.md.

Acknowledgements

• Fine-Grained Head Pose Estimation Without Keypoints (CVPR 2018) by Ruiz, Nataniel and Chong, Eunji and Rehg, James M.
• Head pose estimation by TensorFlow and OpenCV by Yin Guobing

concentration-detection's Issues

运行时报错，请教如何处理

运行时报错，请问下如何解决，能否提供一下requirements.txt，谢谢。
Traceback (most recent call last):
File "concentration_estimation.py", line 178, in
main()
File "concentration_estimation.py", line 65, in main
box_process.start()
File "C:\Users\xxx.conda\envs\tensorflow-gpu\lib\multiprocessing\process.py", line 105, in start
self._popen = self._Popen(self)
File "C:\Users\xxx.conda\envs\tensorflow-gpu\lib\multiprocessing\context.py", line 223, in _Popen
return _default_context.get_context().Process._Popen(process_obj)
File "C:\Users\xxx.conda\envs\tensorflow-gpu\lib\multiprocessing\context.py", line 322, in _Popen
return Popen(process_obj)
File "C:\Users\xxx.conda\envs\tensorflow-gpu\lib\multiprocessing\popen_spawn_win32.py", line 65, in init
reduction.dump(process_obj, to_child)
File "C:\Users\xxx.conda\envs\tensorflow-gpu\lib\multiprocessing\reduction.py", line 60, in dump
ForkingPickler(file, protocol).dump(obj)
TypeError: can't pickle cv2.dnn_Net objects

Traceback (most recent call last):
File "", line 1, in
File "C:\Users\xxx.conda\envs\tensorflow-gpu\lib\multiprocessing\spawn.py", line 105, in spawn_main
exitcode = _main(fd)
File "C:\Users\xxx.conda\envs\tensorflow-gpu\lib\multiprocessing\spawn.py", line 115, in _main
self = reduction.pickle.load(from_parent)
EOFError: Ran out of input

what's the version

I want to run this project in my own computer, and what's the version of the OpenCV and TensorFlow in this project?