ブロックサイズを可変としたDCTによる高精細画像の圧縮方式
第1報 ブロックサイズを可変とした圧縮方式
1996 年 33 巻 1 号 p. 36-45
詳細
1996 年 33 巻 1 号 p. 36-45
This paper deals with the problem of compression of high-resolution images using the discrete cosine transform (DCT). It is well known that DCT is a very efficient transformation for compressing digital images. In particular, DCT with 8×8 block size is adopted in the JPEG algorithm. Since the characteristics of an image vary from place to place, it seems to be more efficient to adopt variable block sizes based on an appropriate index for expressing the features. In that case, the index must be simple and suitable for judging the characteristics of images. First, we analyze some properties of the compressed high-resolution images using DCT with block sizes of 4×4, 8×8 and 16×16, respectively. As high-resolution images, 6 pieces of SCID (Standard Colour Image Data) images are used for simulation. Contrary to the widely believed properties, it will be shown that larger block size is not always better than smaller one. In some cases, the block size of 4×4 may be the best. Next, we present a simple and useful index for judging the best block size. This index is very effective for the purpose of image compression. We can determine the best block size using the index from those of 4×4, 8×8 and 16×16. Various kinds of simulation results using the proposed method are compared with those of the conventional fixed block size system. These results show that variable block size system is very useful for the compression of high-resolution images.
