A coding of a multi-valued image with no distortion has became an important problem in the cases of storing many radiographs and other images. This paper describes image coding methods for data compression of a multi-valued image (256 levels) under the constraint to be no distortion.
In this paper arithmetic coding is used, because it is one of coding system realizes the compression with no distortion. The compression rate is lowered near the entropy of Shannon and processing time is increased with the linear order of a number of symbol.
Generally multi-valued image is a gathering of subimages. For example a part of character, a part of figure, a part of picture and so on. So if multi-valued image is divided usefully into subimages. They have same characteristics with respect to information theoretic views and are compressed rapidly to the entropy of subimage.
This paper describes two methods of dividing multi-valued image into subimage using upper plane information. One is lengthwise division method and the other is the method that divides an image into 8 planes. And each method is done in the case of adaptive and non adaptive scheme. Using now method the compression rate is lower than the compression of convention method.
抄録全体を表示