Programs using complex graphics are showing up in virtually every
area of computing applications including games, education, desktop publishing, graphical
design, and most recently the World Wide Web. Although graphics do a great deal to enhance
the usability and visual aesthetics of such applications, they consume prodigious amounts
of disk storage.
When research into image compression began in the late 1970s, most compression
concentrated on using conventional lossless techniques. However, such types of
compression, which included statistical and dictionary methods of compression, did not
tend to perform well on photographic, or continuous tone images. The primary
problem with statistical techniques stemmed from the fact that pixels in photographic
images tend to be well spread out over their entire range. If the colors in an image are
plotted as a histogram based on frequency, the histogram is not as "spiky" as
one would like for statistical compression to be effective. Each pixel code has
approximately the same chance of appearing as any other, negating any opportunity for
exploiting entropy differences (Nelson 349).
By the late 1980s, extensive research pushed the development of lossy compression
algorithms that take advantage of known limitations of the human eye. Such algorithms play
on the idea that slight modifications and loss of information during the
compression/decompression process often do not affect the quality of the image as
perceived by the human user. Eventually, the JPEG continuous tone image compression
specification was standardized, named after the Joint Photographic Experts Group, the
standards group consisting of members from both the CCITT and the ISO that wrote the
specification. The JPEG specification includes separate lossy and lossless algorithms;
however, the lossless algorithm is rarely used due to its poor compression ratios. Thus,
when one mentions JPEG compression, it can almost be assumed that the reference is being
made to the lossy algorithm, or the JPEG baseline algorithm. The baseline
algorithm, which is capable of compressing continuous tone images to less that 10% of
their original size without visible degradation of the image quality, is detailed below.
