Ohnishi Lab.
Lab. of Biological Information Engineering
Introduction of Research

| Overview | Brain Analysis | Computer Vision | Computer Audition | Natural Language | Rehab. Eng. |

Brain Analysis
Living organisms, in particular represented by humans, have various features that we should take account of. These include:
  1. the excellent perception of seeing, listening, and touching,
  2. Self organization (learning) for adapting to surrounding environment.
  3. integration of various sensory information
  4. active and purposive observation tightly related with motion/behavior,
  5. high level function such as speech generation, comprehension and problem solving.
Analysis and modeling of these biological features not only study the scientific truth, but also contribute to the realization of these biological functions in engineering fields.
Based on the physiological and psychological knowledges, we are engaged in modeling and simulation of visual and auditory processing.

Research theme
  • Model of Visual Organization
    Grouping of optical flows, and Generation of organized structure (form) based on Gestalt's laws.
  • Model of Subjective Contour Generation
    Subjective contours are one of important illusions for understanding human visual information processing. An algorithm based on Gestalt's perceptual organization is proposed for generating subjective contours automatically from given figures.
  • Transparency Perception and its Influence to other Cues for 3-D Reconstruction
    Perception of transparency is one of important cues for objects' overlapping. We are studying how to detect transparency in a image and identify a translucent object, and interactions between transparency perception and other depth cues.
  • Analysis of Binocular Gazing Positions
    We measure and analyze binocular gazing positions to investigate human visual information processing, especially determination of next gazing point and the process from retinal image capturing to spatial recognition.