Human Perception Lab


  • Philip J. Kellman, Ph.D.
  • Philip J. Kellman
  • Principal Investigator
Download CV

Research Interests

  • Visual perception of objects, space, and motion
  • Perceptual learning and development.
  • Applications of perceptual learning, visual cognition, and adaptive learning to education, skill acquisition and educational technology.
  • Human factors applications of perception and cognition research, e.g., perceptual and cognitive processes in aviation and driving.

Research in our laboratory centers on human perception and cognition, with an emphasis on visual perception of objects, space and motion. Our work involves psychophysical experimentation and computational modeling aimed at understanding how perceivers extract information from their environment and derive representations of objects, the layout of environments, and events occurring within it. Focal points include object and surface perception from information that is fragmentary across space and time, visual perception of objects and space by moving observers, perception of shape, and interactions among object, space and motion perception.

We also study perceptual learning and its relations to other aspects of learning, expertise and cognition. Research in recent years is helping to broaden ordinary conceptions of learning and instruction that focus on declarative and procedural aspects. These are important but incomplete in that they do not include the dramatic effects of experience in any learning domain in changing the way information gets encoded (perceptual learning). As studies of expertise have revealed, domain-specific attunements of information extraction, discovery and encoding of new relational structure, and improvements in fluency of information pickup are crucial components of becoming good at anything. Our work seeks to understand these perceptual learning effects, and we also develop and test methods of training and accelerating expert information extraction and pattern recognition skills. Synergistic with this work are novel adaptive learning algorithms that improve the efficiency of learning in any domain (factual, procedural, perceptual learning, or combinations of these). These efforts in computer-based learning technology have already produced dramatic advances in learning, retention, and transfer in diverse domains, such as aviation training, mathematics and science learning, and medical learning. We also seek deep models of high-level perceptual learning that capture the abstract properties of human perception and perceptual learning.