Computer vision: a unified, biologically-inspired approach
This unique volume is a comprehensive, self-consistent coverage of one approach to computer vision, with many direct or implied links to human vision. The book is the result of many years spent by the author in research into the limits of human visual performance and the interactions between the observer and his environment. A wide-ranging and largely novel approach to computer vision is described. The treatment starts with a summary account of important aspects of human visual function. This is then followed by a progressive development of the computer image processing, from sub-pixel fragmentary edge determination to optical flow field analysis, local and global stereo analysis, colour imagery, edge-based region segmentation, perceptual texture segmentation and high fidelity contour form analysis. This treatment is considerably different from other more publicised approaches, and is highly recommended for those seeking a dynamic new approach to computer vision.
What people are saying - Write a review
We haven't found any reviews in the usual places.
Some General Concepts
Combined First and Second Difference Edge Sensing
17 other sections not shown
1st difference 2nd difference adjacent algorithms approximately array basic blur carried centre Chapter chrominance colour colour constancy components computer vision considered corners curvature disc discussed edge data edge strength effects energy errors extraction feature flow vector fovea fragmentary edge fragmentary profile frames Gaussian Gaussian blur global grey level hexagonal high fidelity human vision human visual image processing input image interactive interpolation Laplacian luminance matrix mean measure motion vector neighbourhood neural object operators optical flow orientation difference original image orthogonal output Overington pair peak perception pixel point spread function possible practical preprocessor problem Proc progressive receptive fields receptor region boundaries region segmentation resolution retinal sampling scale scene sensing sensor shown in Fig signal simple spatial frequency square standard deviation statistics stereo disparity stimulus studies sub-pixel temporal texture threshold tion trends typical vernier position vision systems visive visual system whilst window analysis zero crossing