Computational Modelling of the Human Visual Cortex
- LecturerDr. Dacheng TAO (NANYANG Assistant Professor, Nanyang Technological University)
Host: Dr. Hong-Yuan Mark Liao - Time2010-07-13 (Tue.) 14:00 – 16:00
- LocationAuditorium 106 at new IIS Building
Abstract
Abstract:
When a photon impinges on the retina, the light signal is
transduced into an electrical signal. Afterward, the
electrical signal is conveyed in turn to lateral geniculate
nucleus (LGN), visual cortex, inferior temporal cortex (ITC),
prefrontal cortex (PFC), medial temporal lobe (MTL) and
other brain areas, and thus we can perceive the visual world,
recognise objects, identify our friends, analyse events,
remember every happy time, and guide our actions toward
specific objects. All these tasks are very tough challenges
for current computers, so it essential to understand the
biological visual system that can guide our research on
computational vision. In this talk we discuss our recent
work on computational modelling of the ventral path way.
This is a further step of the recent MIT’s follow up to
Hubel and Wiesel’s feed-forward hierarchy model published
in 1959. We particularly considered the feedback process
between the visual cortex and MTL that consolidates the
memory and assists the visual recognition. Our preliminary
experimental results on object recognition support our model.
Surely, this model is general and can be adapted to model
the dorsal path way for motion analysis.
Short-Bio:
Dacheng Tao received the Ph.D degree from the University of
London. Currently, he is a NANYANG Assistant Professor in
the Nanyang Technological University, a Research Associate
Fellow in the University of London, a Visiting Professor in
the Xidian University and a Guest Professor in the Wuhan
University. He works on computational neuroscience,
biologically inspired models, statistics and their
applications in computational vision and video surveillance.
He has authored more than 150 scientific articles at top
venues including IEEE T-PAMI, T-IP, T-KDE, NIPS, KDD, ICDM,
and AISTATS with best paper awards. He is an associate
editor of IEEE Transactions on Knowledge and Data
Engineering (T-KDE) and Elsevier Neurocomputing. He has (co-
)chaired more than 30 times for special sessions, invited
sessions, workshops, panels and conferences. He has served
for more than 110 major international conferences including
ICDM, KDD, CVPR, ICCV, and ECCV, and more than 50
prestigious international journals. He is a member of IEEE
and an IEEE TC on Cognitive Computing.