Compression des images numériques par la technique des ondelettes

Abstract

The objective of this work is to study the of digital still images to grayscale and color with wavelet technique , specifically JPEG 2000 (Joint Expert Group) still image compression standard based on wavelets . This new method, recently proposed in the literature, has received special attention of many researchers in recent years. The JPEG compression standard based on the DCT (Discrete Cosine Transform) is very successful and remains a method of image compression very efficient in the field of Internet and digital cameras. For high compression, the image quality degrades rapidly. In addition, JPEG offers very little resistance to errors due to variable length coding type Huffman used. JPEG2000 had as main objectives to meet the diverse needs of newapplications and provide a clear improvement over JPEG in terms of quality and resistance to transmission errors. Comparative performance tests of this method in terms of and quality of the reconstructed images with the JPEG compression method (Baseline) in mode losses were done with varying test images (in grayscale , color and Medical) underMatlab7. The results show that JPEG2000 is better than the JPEG standard and more tests were also done on the new features provided by JPEG2000 as the of Interest (ROI), coding with and without loss image with a single algorithm, decoding the image with different resolution levels or with different levels of quality and also with different levels of and quality at the same time.