engleski

Features of Wavelet Filters in Still Image Compression

The paper discusses the characteristics of wavelet filters used for transformation and compression of various image types. The basis of wavelet transform as well as features and classification of filter family is given. The results and comparison of filters for various image types in example of Lena, Zebra and Baboon images is given. The results show that the content of an image and filter used for transformation and compression is very important for compressed image quality. The aim of this work is to create a palette of functions (filters) for implementation of wavelet in still image processing and to emphasize the advantage of this transformation relating to today's methods such is DCT in JPEG. Here are given the samples of the most used filters and comparison of these filters for various images and sizes. Filters taken in our test are some of the most used Haar filter (as the basis), orthogonal Daubechies' and Battle-Lemarie filters such is Least Asymmetric, Shannon and Coiflet, and symmetric Biorthogonal filter. All these filters give various performances for images of different content. Objective and subjective picture quality characteristics of images coded using wavelet transform with different filters are given. The comparison between JPEG coded picture and the same picture coded with wavelet transform is given. For higher compression ratios it is shown that wavelet transform has better S/N than JPEG.

wavelet; filters; image compression; signal-to-noise ratio

nije evidentirano