Auto-evaluating Image Retargeting Algorithms
Event date:  October 4, 2018 4:00 PM  to 5:00 PM

There are numerous situations where images need to be auto-adapted into different sizes and compositions. For instance, the increasing diversity of visual display devices (especially mobile ones) has demanded for effective image adaptation (retargeting) for a given aspect ratio; automatic (or semi-automatic) image editing is required for Photoshop (and other similar tools), native advertising, webpage content creation, post-processing of TV/movies, VR/AR scenarios, etc; there will be also future possibilities for reducing the amount of visual data  transmitted over the wireless network (with low bit-rates) while ensuring the required accuracy for a specific AI task. To date, there is not a single image retargeting method that works consistently well for all visual content, and this call for effective, objective image retargeting quality assessment (IRQA) to be developed for on-line benchmarking and then selecting the best retargeting scheme dynamically at servers; furthermore, IRQA facilitates the advancement of image retargeting techniques themselves. In this talk, a new IRQA methodology is presented, being able to predict the perceived retargeting quality by proper evaluation of salient content loss and visual distortion. Its key contribution lies in dense correspondence (DC) estimation and multi-level feature (MLF) fusion: the former makes IRQA truly full-reference assessment for the first time, while the latter enables better agreement with human appreciation and adjustment.


Lin Weisi is an active researcher in image processing, perception-based signal modelling and assessment, video compression, and multimedia communication systems. In the said areas, he has published 180+ international journal papers and 230+ international conference papers, 7 patents, 9 book chapters, 2 authored books and 3 edited books, as well as excellent track record in leading and delivering more than 10 major funded projects (with over S$7m research funding). He earned his Ph.D from King’s College, University of London. He had been the Lab Head, Visual Processing, Institute for Infocomm Research (I2R). He is a Professor in School of Computer Science and Engineering, Nanyang Technological University, where he served as the Associate Chair (Graduate Studies) in 2013-2014. 





University of Konstanz, Universitätsstr. 10, Konstanz
Z 613

Unfortunately, the talk will not be transmitted to Stuttgart and Tübingen.