In this repository we provide code of the paper:

Adaptive Compressed Sensing for Real-Time Video Compression, Transmission, and Reconstruction

Yaping Zhao, Qunsong Zeng, Edmund Y. Lam

paper link: https://ieeexplore.ieee.org/abstract/document/10302598

The videos used in our experiments are available at OneDrive.

Please extract video.zip to video folder.

Simply run:

python main.py

Navigate to the video folder. You will see files video*.mp4 as the original videos, and meas*.mp4 as the compressed videos.

In our paper, we adopt the classical GAP-TV algorithm to reconstruct videos. For higher reconstruction quality, we recommend DEQSCI.

###Adjust the Compression Rate Modify the parameter cr, which is the compression rate, in Line 53 of the file main.py.

......
### compress video frames into measurements

    output_path = './video/meas%d.mp4' %i
    img_num = len(os.listdir(folder_name))
--> cr = 8 # compression rate
    meas_num = img_num // cr
......

Cite our paper if you find it interesting!

@INPROCEEDINGS{10302598,
  author={Zhao, Yaping and Zeng, Qunsong and Lam, Edmund Y.},
  booktitle={2023 IEEE 10th International Conference on Data Science and Advanced Analytics (DSAA)}, 
  title={Adaptive Compressed Sensing for Real-Time Video Compression, Transmission, and Reconstruction}, 
  year={2023},
  volume={},
  number={},
  pages={1-10},
  doi={10.1109/DSAA60987.2023.10302598}}