Many visual illusory phenomena are known in relation to senses of color, size, and depth, and to understand their mechanisms is to understand the basic function of the senses. Our three research groups each use their own methods to engage in mathematical modeling of visual illusions, and based on mutual exchange, link that work to a variety of applications.

The Sugihara Group is exploring the sensory functions of color, size, and depth from the perspective of models of solid visual illusions, with applications in developing methods of media expression capable of conveying the magnificence of the natural world, configuring standards for regulating advertisements that exaggerate width and depth of space, and alleviating traffic congestion by avoiding misapprehension of incline on sloping roads.

The Arai Group is seeking to explore visual illusions by modeling neurological function at each level of the brain. The group is also developing technology to control increase and decrease in the strength of visual illusion.

The Yamaguchi Group is modeling the gap in visual information between signal content and meaning content for humans, to develop a method for evaluating quality of visual information. Evaluation of the sensory quality of such things as color and size has applications in improving the quality of image-based dialogue between humans and machines.

Mathematical models of visual illusions are built for situations in which a solid structure has been identified from a retinal image, the strength of visual illusion is extracted from the model, and a method of control is developed. That is then used to improve safety and cultural richness. In particular, by focusing on such visual illusions as the incline on a slope, depth distance, subjective contour lines, impossible solids, and impossible motion, and with identification methods used by computers as a reference, computational models of human visual illusions can be built and visual illusory effect quantitatively extracted. The strength of visual illusion can then be explored to see how reliant it is on environmental conditions, and based on that, technology can be built to minimize or maximize the strength of visual illusion. Such control technology is not only useful for society, but methods for mathematical modeling capable of expressing flexible and high performance human senses and awareness, and robust computational methods to operate them, can also be systemized through such research activities.

Mathematical models of vision are built using perspectives from neuroscience, brain science, and sensory psychology. The method for determining the appropriateness of the mathematical models thus created is as follows: a computer into which the mathematical model has been loaded is investigated to see whether, like a human, it manifests visual illusions. For that purpose, a visual illusion is used that is thought to be associated with a particular area of the brain. Classic visual illusions are used, but unique new visual illusion figures are also conceived and simulated.

Conversely, computational algorithms are conceived to enable simulation of visual illusions invented by members of the Arai Group, or of known classic visual illusions for which manifestation mechanisms are yet to be elucidated, and the unknown mechanisms for processing visual information in the brain are inferred in reverse to elucidate mechanisms that give rise to visual illusions.

In the later stages of research, mathematical tools developed and used for research into visual illusions and vision, such as wavelet frames, are delved into mathematically, generalized, and applied to various fields.

These plans are originally proposed by Hitoshi Arai, and are now executed by Arai Group.

Research will be further undertaken into the relationships between art and the visual and mathematical models created.

A range of physiological responses including EEG, eye tracking, and EDA are measured so that unique measurements which manifest brain and cognitive mechanisms associated with visual illusory phenomena can be analyzed. This serves to elucidate conscious and unconscious biological reactions to visual illusory phenomena, and assists in quantifying visual illusions. We particularly advance analysis of the brain mechanisms for vision that are closely associated with visual illusions. In parallel, making use of visual illusory phenomena, we investigate the relationship between biological reactions and strangeness in vision, and knowledge can be gathered on biological reactions and measurements that reflect the strangeness. This knowledge then becomes the basis for developing methods to evaluate the sensation quality of images, and attempts can also be made to apply the knowledge to methods for generating images.The intent is to make it possible to evaluate and create images.