Abstract:
Inspired by Shannon's work on estimating the entropy of a language, we experimented with a framework for image compression comprising one human describing images using text instructions to another, who is tasked with reconstructing the original image to the first human's satisfaction. These image reconstructions were then rated by human scorers on the Amazon Mechanical Turk platform and compared to reconstructions obtained by existing image compressors. While this setup lacks certain components typical of traditional compressors, the insights gained from these experiments offer a perspective on the potential for substantial improvements over current approaches to lossy image compression.
Joint work with Ashutosh Bhown, Soham Mukherjee, Sean Yang, Shubham Chandak, Irena Fischer-Hwang and Kedar Tatwawadi.
Bio:
Tsachy Weissman has been on the faculty of the Electrical Engineering department at Stanford since 2003, where he enjoys activities such as research and teaching the science of information. He has served and still does on editorial boards for scientific journals, technical advisory boards in industry, and as founding director of the Stanford Compression Forum. His favorite gig to date was being an advisor to the HBO show “Silicon Valley” until he was terminated when it was realized his students are more creative and reliable consultants. He hates writing about himself in the third person.