ABSTRACT
With the rapid advancement in imaging technology, color images are more easily available and used in different applications than ordinary black and white or gray scale images. Increasing use of color has led to formulation of new challenges and problems in image processing. For example, various image capturing and display devices have their own limitations on data acquisition and rendition. Hence, different techniques and algorithms have been advanced
to overcome their limitations and to preserve color information as images are exchanged between devices. Mostly, these algorithms are developed in the spatial domain. However, a few techniques have also been reported for processing color images in the compressed domain, in particular in the block discrete cosine transforms (DCT) domain. In this chapter we discuss some of the techniques and problems of color processing in the block DCT space. First, we briefly introduce the fundamentals related to color processing, which is followed by a discussion on different approaches in the compressed domain.
Newton’s pioneering experiments on sunlight revealed the fact that the color of light depends upon its spectral components. Later, Grassmann [54] and Maxwell [97] laid the mathematical foundation of representation of a color factored into three independent variables. Grassmann stated the law of color matching [159], and Maxwell demonstrated that any additive color mixture could be matched by appropriate amount of three primary stimuli [98]. This is referred to as trichromatic generalization or trichromacy. For a long time, psychophysical experiments [159] were carried out to determine color matching functions (CMFs), which form a set of three functions providing numerical specification of color in terms of three coordinates or tristimulus values. Hence, it is the presence of different wavelengths of optical light and their relative strength that determine the sensation of color. This characteristics of a light source is expressed in terms of spectral power distribution (SPD). According to our perception, we attribute these three factors of color to the following.
