Compression of image

The compression of image is an application of the Data compression on digital images. This compression has as a utility to reduce the redundancy of the data of an image in order to be able to store it without occupying much space or transmitting it quickly.

The compression of image can be carried out with loss of data or without loss . Compression without loss is sometimes preferred for artificial images such as the diagrams, the technical designs, the icons or the cartoons. It is because of the fact that the methods of compression with loss, when in particular they are used with low binary debits, present defects of compression. Methods of compression without loss can also be preferred to keep a high degree of accuracy, such as for medical sweepings, or digitalizations of images intended for filing. The methods with loss are particularly appropriate to the normal images such as photographs in the applications where a minor loss of fidelity (sometimes unperceivable) is acceptable to carry out a substantial reduction of binary debit.

The most important methods of compression of image without loss are:

  • method of the coding of the repetitions,

  • the Coding of source,
  • algorithms with adaptable dictionary such as LZW.

The most important methods of compression with loss are:

  • the reduction of the space of the colors to the most frequent colors in an image. The selected colors are indicated in the palette in the heading of the compressed image. Each pixel indicates just a reference on a color in the palette.
This method can be combined with the Dithering to scramble the borders of color.
  • the Under-sampling of the chrominance. This method profits owing to the fact that the eye perceives the luminosity more vigorously than the color, while dropping at least half of information from chrominance in the image.
  • the Coding by transformation. It is generally the most used method. The Transformed into cosine discrete and the transformation by ondelettes are the most popular transformations. Coding by transformation includes/understands the application of the transformation to the image, followed by a quantification and a entropic coding.

See too

Measurements of quality:

Random links:Proteus (kind) | Raimon Panikkar | Chaetodon sanctaehelenae | Faberrebe | Name-of-father | Convertisseur_tournant