Multilevel Image Coding Using a Two-Component Method
Original Publication Date: 1988-Dec-01
Included in the Prior Art Database: 2005-Feb-15
This process enables encoding of a time-independant multilevel monochrome image (e.g., freezed frame video or digitized black and white photograph) with very low bit rate (< 1 bit/pixel) and very low distortion. This process is based on two-component image coding principles. The original image is first digitized into an array of integers quantifying the luminance signal in each sampling point (pixel (x,y)) of the image using 256 gray levels (8 bits/pixel). The corresponding data flow I(x,y) is then compressed through a 3-step process involving: 1.Segmentation of digitized image into closed homogeneous regions 2.Plateau component descriptor determination and coding. This step determines an approximation Ip(x,y) of luminance signal I(x,y) within each closed region; Ip is either an horizontal plane or an oblique plane. 3.