engleski

Wavelet downsampling in image compression

Wavelet downsampling is a method of data compression. However, it is very possible that part of information might be lost when such method is used. These loses should be compensated by interpolation in reconstruction algorithm. Simple method of interpolation is used in this paper. Reconstruction algorithm was tested. The goal is to find out how much of downsampling specific data can take without undermining the purpose. For example, how much can be lost in medical image without change in diagnosis. The study can be repeated in different applications. We used the simplest wavelet - Haar. The results show that X-ray pulmonary images can be downsampled up to 10th level without change in diagnosis. In computer memory space, that means that image from size of approximately 4, 5 MB can be reduced to the size of approximately 35 KB or 128 times.

wavelet; image processing; wavelet compression

nije evidentirano