Other Free Encyclopedias » Online Encyclopedia » Encyclopedia - Featured Articles » Contributed Topics from F-J

JPEG-LS

context values pixel image

Definition: JPEG-LS is an image compression standard (ISO/IEC-14495-1 / ITU-T Rec. T.87) designed to provide effective lossless and near lossless compression of continuous-tone, gray scale and color still images.

Near lossless , in this context, means a guaranteed maximum error between the original image data and the reconstructed image data. One of its main potential driving forces is the compression of medical images without any quality loss.

The core algorithm behind JPEG-LS is called LO w Co mplexity LOssless Compression for Images (LOCO-I), proposed by Hewlett-Packard. LOCO-I builds upon a concept known as context modeling . The main idea behind context modeling is to take advantage of the structure in the input source, modeled in terms of conditional probabilities of what pixel values follow from each other in the image (Figure 1).

LOCO-I has three main components:

  • Predictor: uses a fixed predictor, whose output will be: a (when there is a vertical edge to the left of the current pixel location); b (when there is a horizontal edge above the current pixel location); or a + b – c when the neighboring pixels have comparable values.
  • Context model: the context model that conditions the current prediction error ( residual ) is indexed using a three-component context vector whose values represent the local gradients around the reference pixel. These values are quantized and the vector is mapped into an integer in [0, 364].
  • Coder: the error residuals are coded using adaptively selected codes based on Golomb codes, which are optimal for sequences with geometrical distributions.

For more information on JPEG-LS, including code, please refer to.

Jubayr, Adil al- (1962–) - PERSONAL HISTORY, BIOGRAPHICAL HIGHLIGHTS, PERSONAL CHRONOLOGY:, INFLUENCES AND CONTRIBUTIONS, THE WORLD’S PERSPECTIVE, LEGACY, CONTEMPORARIES [next] [back] JPEG 2000 Image Coding Standard - Abstract, Introduction, The JPEG 2000 Coding Algorithm, The Image Transform, Quantization, Entropy coding

User Comments

Your email address will be altered so spam harvesting bots can't read it easily.
Hide my email completely instead?

Cancel or