Mathematical Methods in Computer Vision [Paperback]

This volume comprises some of the key work presented at two IMAWorkshops on Computer Vision during fall of 2000. Recent years haveseen significant advances in the application of sophisticatedmathematical theories to the problems arising in image processing.Basic issues include image smoothing and denoising, image enhancement,morphology, image compression, and segmentation (determiningboundaries of objects-including problems of camera distortion andpartial occlusion). Several mathematical approaches have emerged,including methods based on nonlinear partial differential equations,stochastic and statistical methods, and signal processing techniques,including wavelets and other transform theories.Shape theory is of fundamental importance since it is the bottleneckbetween high and low level vision, and formed the bridge between thetwo workshops on vision. The recent geometric partial differentialequation methods have been essential in throwing new light on thisvery difficult problem area. Further, stochastic processes, includingMarkov random fields, have been used in a Bayesian framework toincorporate prior constraints on smoothness and the regularities ofdiscontinuities into algorithms for image restoration andreconstruction.A number of applications are considered including optical characterand handwriting recognizers, printed-circuit board inspection systemsand quality control devices, motion detection, robotic control byvisual feedback, reconstruction of objects from stereoscopic viewand/or motion, autonomous road vehicles, and many others.This volume comprises some of the key work presented at two IMA Workshops on Computer Vision during fall of 2000. Recent years have seen significant advances in the application of sophisticated mathematical theories to the problems arising in image processing. Basic issues include image smoothing and denoising, image enhancement, morphology, image compression, and segmentation (determining boundaries of objects?including problems ofl#O