Speakers

Dr Anastasia Mozhaeva

Topic

The human visual system processes enormous amounts of visual information while selectively prioritising what is most relevant for perception and action. This seminar explores how experimentally measured characteristics of human vision can be translated into computational models for more efficient image processing and artificial intelligence. I will present our research on spatial, temporal, and peripheral visual perception, and show how these findings can be used to reduce perceptually redundant visual information before transmission or AI inference. Applications will include perception-aware video processing, foveated and wide-angle visual systems, virtual reality, embedded AI, and autonomous platforms. The seminar will also discuss how biologically inspired visual processing can reduce computational and communication demands while maintaining, and in some cases improving, the performance of existing computer vision algorithms.

About the Speakers

Dr Anastasia Mozhaeva is a Senior Lecturer in Computing at the Eastern Institute of Technology and holds a PhD in Electronics from the University of Waikato. Her research combines human visual perception, image and video processing, embedded artificial intelligence, and machine vision. She develops perception-aware approaches that use experimentally measured characteristics of human vision to improve the efficiency of AI and visual communication systems. Her research extends to embedded and autonomous platforms, including drones and wide-angle visual systems. She will also be joined by Steve Rogers, a former EIT student who is now a lecturer and continues his research under her supervision.