This chapter discusses the image enhancement by means of gray-scale transformations, smoothing, and shading compensation. The chapter also discusses the concept of detail enhancement, enhancement of the local contrast, line enhancement, and binary image enhancement. Most of the enhancement in gray-scale transformations can be described by point operations. These operations are transformations of the gray scale, in which the current pixel's new gray value is a function only of the original gray value, but not of the gray values of some neighboring pixels. The operator is completely specified by the input/output gray-value characteristic, also called the gray-value transformation function. If the gray-value deviations with respect to the flat model inside a homogeneous region are of an additive nature, smoothing can be accomplished by means of a linear minimum square error (LMMSE) technique. This simple and efficient technique lends itself to local adaptation, and it has been used successfully in several applications. The LMMSE technique determines the least-mean-square estimation at a point under the constraint of being a linear combination of the current gray value and of the local expected value, approximated by the local mean. © 1995 Academic Press Inc.
CITATION STYLE
Zamperoni, P. (1995). Image Enhancement. Advances in Imaging and Electron Physics, 92(C), 1–77. https://doi.org/10.1016/S1076-5670(08)70006-5
Mendeley helps you to discover research relevant for your work.