Stanford
Stanford Compression Forum

Human Compression

Humans are awesome*

Ashutosh Bhown, Soham Mukherjee, Sean Yang, Shubham Chandak, Irena Fischer-Hwang, Kedar Tatwawadi, Tsachy Weissman
 

Abstract: Lossy image compression has been studied extensively in the context of typical loss functions such as RMSE, MS-SSIM, etc. However, it is not well understood what loss function might be most appropriate for human perception. Furthermore, the availability of massive public image datasets appears to have hardly been exploited in image compression. In this work, we perform compression experiments in which one human describes images to another, using publicly available images and text instructions. These image reconstructions are rated by human scorers on the Amazon Mechanical Turk platform and compared to reconstructions obtained by existing image compressors. In our experiments, the humans outperform the state of the art compressor WebP in the MTurk survey on most images, which shows that there is significant room for improvement in image compression for human perception.

* lossy image compressors

Arxiv: Link

Data:
  • Original images (*_original*), human reconstructions (*_final*) and WebP reconstructions (*_webp*): Link
  • Chat transcripts (*_verified.txt) and the corresponding bzip2 compressed files (*_verified.txt.bz2): Link
  • MTurk survey results for human reconstructions (image_name.csv) and for WebP reconstructions (image_name_webp.csv): Link
  • Code and other parameters: Link
  • Statistical analysis of results: Link, Github

Twitter: