ABSTRACT:
Joint work with Michele Covell, Chris Bregler and Gerald McRoberts
I would like to talk about several applications of signal processing to human auditory perception. The human cognitive system is the last great frontier for research--I would like to explore our understanding of perception by building systems to model and recognize human behaviors. Over the last few years I have explored many areas of human understanding and synthesis, which I would like to share with you. Topics I will discuss include human emotion recognition (BabyEars), measuring the synchronization between video and audio of talking faces (FaceSync), manipulating the speed of speech in a perceptually realistic manner (Mach1) and synthesizing talking faces (Video Rewrite).
Many audio and video examples will be presented. Bring your own popcorn.
About Malcolm Slaney:
Malcolm Slaney currently is a research scientist at IBM's Almaden Research Center. Prior to that, he was a research scientist at Paul Allen's Interval Research Corporation, and a research scientist at Apple Computer. His background is in electrical engineering, signal processing, and auditory psychophysics. He is co-editor with Steve Greenberg of "Computational Models of Auditory Function (2001), and has done a great deal of interesting work on visual representations of auditory signals including the correlogram and computational auditory scene analysis.