Space-Variant Image Restoration

摘要: Publisher Summary This chapter offers information on space-variant image restoration. The inhomogeneity assumption introduced in this allows the edges to be directly taken into account by model. A stochastic image-generating process is so obtained describing gray level discontinuities a space-varying model, where only edge location needed. Consequently, optimal restoration procedure guaranteed corresponding nonstationary Kalman filter. Blurring intrinsically avoided because at any pixel estimate using carried neighboring pixels belonging convex set contained same subregion. can an edge-detector operator, whose main feature should robustness with respect noise. In discussing filtering, brief approach filtering of noisy signals presented. discussion sampled as all data processing currently performed means computers; moreover, mathematics involved greatly simplified. presents constitutive equation, discusses realization image, explains deblurring, and on.