Items related to Cepstral Filtering on a Columnar Image Architecture:...

Cepstral Filtering on a Columnar Image Architecture: A Fast Algorithm for Binocular Stereo Segmentation (Classic Reprint) - Softcover

Yehezkel Yeshurun

 
9781332111435: Cepstral Filtering on a Columnar Image Architecture: A Fast Algorithm for Binocular Stereo Segmentation (Classic Reprint)

Synopsis

This book presents a novel algorithm for object segmentation in stereo vision, designed by the author to mirror principles of the human visual cortex. The algorithm analyzes a pair of stereo images through a windowed cepstral filter operating on a columnar image architecture inspired by ocular dominance columns. The result is a one-step measurement of disparity, and is resistant to disturbances including blur, size differences, and intensity changes. By explaining and discussing the algorithm in relation to the limitations and advantages of human stereo vision, the author presents a computational solution to a problem which has been elusive in the field of computer vision. Furthermore, the author suggests that the algorithm provides a rationale for the existence of columnar systems in the human brain, both with regard to ocular mixing, and to other modalities. The book will be of interest to researchers in computational stereo and vision science, as well as scholars and students of cognitive science, neuroscience, and artificial intelligence.

"synopsis" may belong to another edition of this title.