Deep-NC: A secure image transmission using deep learning and network coding

作者：

Highlights：

• Integrating network coding with image super resolution can conceal and reduce image file size.

• Eavesdropper cannot recover the original image without knowledge of reference image.

• Per-pixel coding achieves a better quality than per-image coding.

• Per-image coding requires a lower computational complexity with a lower bandwidth.

• Images can be downscaled to a much lower resolution to save the transmission bandwidth.

摘要

•Integrating network coding with image super resolution can conceal and reduce image file size.•Eavesdropper cannot recover the original image without knowledge of reference image.•Per-pixel coding achieves a better quality than per-image coding.•Per-image coding requires a lower computational complexity with a lower bandwidth.•Images can be downscaled to a much lower resolution to save the transmission bandwidth.

论文关键词：Image communication,Deep learning,Super-resolution,Network coding

论文评审过程：Received 2 February 2021, Revised 5 July 2021, Accepted 10 September 2021, Available online 20 September 2021, Version of Record 2 October 2021.

论文官网地址：https://doi.org/10.1016/j.image.2021.116490