Research Topics
Human sensory processing
The human sensory system has marvelous abilities to recognize a wide variety of things, stuffs and events in the real world, ranging from basic sensory attributes such as motion and shape of the object, time and space, material parameters, ecological values, to human emotional states. We aim at understanding the computational theory, as well as the representations and algorithms, of the neural sensory computation. The scientific approach we take is mainly psychophysics, a methodology to infer the brain's sensory processing from the behavioral responses and subjective perception for a variety of controlled stimuli, in combination with the cutting-edge media information technologies including machine vision and computer graphics. The scope of this study is not limited to vision, but also includes touch, hearing, multisensory integration, and metacognition. We pursue deeper understanding of human cognitive processing through comparison with artificial neural networks that show human-competitive performance and brain-like internal representations.
Perceptually based media technologies
A key idea to develop innovative media technologies often lies in how to cheat the human cognitive system in clever ways. For example, it has been shown that by making use of human perceptual properties, technologies for visual display and augmented reality could improve in efficiency in comparison with physical reproduction of the sensory input signals, or achieve such functions impossible to realize with physically based methods. We plan to continue development of innovative perceptually based media technologies, making use of the achievements in cognitive neuroscience for information technologies.
Human decision-making and metacognitive monitoring
Humans make decisions based on information from the external environment. Moreover, the human brain recursively monitors its internal representations to detect errors and make flexible behavioral adjustments. These representations in the brain often exhibit specific statistical structures, allowing for efficient processing through heuristic shortcuts. Standing upon this ecological rationality perspective, we investigate the information processing that underlies human decision-making.
