JPEG compression

The acronym JPEG (Joint Photographic Expert Group comes from the meeting held in 1982 of a group of photographic experts, whose main concern was to work on the ways to transmit information (still or animated images). In 1986, the ITU-T developed compression methods intended for fax sending. These two groups joined to create a joint photographic experts group (JPEG).

Unlike LZW compression, JPEG compression is a lossy, , which affords it one of the best compression ratios, despite the slight loss of quality (20: 1 with 25:1 without any significant loss of quality).
This compression method is much more effective for photographic images (comprising many different coloured pixels) than on geometrical images (unlike LZW compression) because the nuance differences due to compression are very visible in the latter.

The stages of JPEG compression are the following:

• Chrominance re-sampling, because the eye cannot distinguish chrominance differences within a square of 2x2 points
• • Image division into blocks of 8x8 points, then the application of the DCT function (Discrete Cosine Transform, which breaks up the image into a sum of frequencies
• Quantification of each block, i.e., a loss coefficient is applied (which makes it possible to determine the size/quality ratio) which will “cancel out” or decrease high frequency values, in order to attenuate the details by intelligently going over the block with RLE coding (in a zigzag pattern in order to remove as many zero values as possible).
• • Image encoding then compression using the Huffman method

File formats that save a flow coded in JPEG are actually called JFIF (JPEG File Interchange Format), but the term is usually deformed into “JPEG file".

It should be noted that there is a lossless form of JPEG coding. Although little used by the data-processing community in general, it is used especially for the transmission of medical images in order to avoid confusing artefacts (purely dependent on the image and its digitalization) with real pathological signs. Compression is thus much less effective (only factor 2).

Last update on Thursday October 16, 2008 02:43:16 PM.This document entitled « JPEG compression » from Kioskea (en.kioskea.net) is made available under the Creative Commons license. You can copy, modify copies of this page, under the conditions stipulated by the licence, as this note appears clearly.