Room 1037 (Emerging Technology Building- ETB)
Dr. Yoonsuck Choe/TAMU/CSE
What is the nature of texture? There are several theories regarding this question, including the texton theory and an extended statistical theory. Although these theories and approaches to texture processing may differ in detail, they share the basic idea that texture is a vision problem. However, given that texture is most directly associated with a 3D surface (e.g., that of a tree bark or a stone), there is a possibility that texture can be fundamentally tactile. Interestingly, tactile receptive fields found in the somatosensory area 3b resemble that of the visual receptive fields, albeit with an important difference. In the tactile receptive fields, in addition to the Gabor-like components, there is an inhibitory component that changes its location dependent on the scanning direction of the tactile patch. Our experiments resulted in tactile response representation outperforming visual response representation in texture tasks. Furthermore, an unsupervised self-organization experiment showed that an identical cortical self-organization model develops receptive fields resembling tactile receptive fields when trained with various texture-like stimuli, while it develops visual receptive fields with natural-scene-like stimuli. These results suggest an intimate link between texture and the tactile modality. We expect our findings to help us better understand the nature of texture, and allow us to develop more powerful texture processing algorithms based on tactile receptive fields.
Bio: Yoonsuck Choe is an associate professor and director of the Brain Networks Laboratory in the Department of Computer Science and Engineering at Texas A&M University. He received his B.S. degree in Computer Science from Yonsei University (Korea) in 1993, and his M.S. and Ph.D. degrees in Computer Sciences from the University of Texas at Austin in 1995 and 2001. His current research areas include connectomics (microscopy instrumentation, computational neuroanatomy and neuroinformatics), computational neuroscience, neural networks, and neuroevolution.