Learned Image Compression and Visual Perception

Lecture / Panel
For NYU Community



Johannes Ballé
Staff Research Scientist, Google


"Learned Image Compression and Visual Perception"


Today, images and videos make up more than 60% of the Internet’s bandwidth, with a growing tendency. Since its emergence about half a decade ago, the field of learned data compression has attracted considerable attention, because it promises to significantly lower this burden. Using machine learning in source coding means faster innovation cycles, as well as better adaptation to novel data modalities (such as AR/VR formats) and to nonlinear measures of visual quality, all of which ultimately result in greater compression efficiency. For example, image codecs can now be end-to-end optimized to perform best for specific types of images, by simply replacing the training set. They may now be designed to minimize a given perceptual image metric, or in fact any differentiable perceptual loss function. In this talk, I will first review the current state of learned data compression, and then discuss the two challenges which need to be solved to replace JPEG on your phone, tablet, laptop, TV, and refrigerator: developing better models of human perception, and lowering computational complexity in the process.

About Speaker

Johannes Ballé (they/them) is a Staff Research Scientist at Google, currently focusing on lossy image compression, information theory and models of visual perception. They defended their master's and doctoral theses on signal processing and image compression under the supervision of Jens-Rainer Ohm at RWTH Aachen University in 2007 and 2012, respectively. This was followed by a brief collaboration with Javier Portilla at CSIC in Madrid, Spain, and a postdoctoral fellowship at New York University’s Center for Neural Science with Eero P. Simoncelli, where Johannes studied the relationship between perception and image statistics. While there, they pioneered the use of variational Bayesian models and deep learning techniques for end-to-end optimized image compression. They joined Google in early 2017 to continue working in this line of research. Johannes has served as a reviewer for top-tier publications in both machine learning and image processing, such as NeurIPS, ICLR, ICML, Picture Coding Symposium, and several IEEE Transactions journals. They have been active as a co-organizer of the annual Challenge on Learned Image Compression (CLIC) since 2018, and on the program committee of the Data Compression Conference (DCC) since 2022.