Image coding with optimal decoding possible afterwards

Abstract

Instead of decompressing a whole image, in practice it is often necessary to decompress a certain part of it. Most methods, upon the request for certain image fragment, in particular a small one, must first decompress the whole image, thereby wasting excess memory resources. Therefore, these methods have limited use, since there might not be enough memory resources to decompress the whole image on the hardware being used. The new coding method SS-SPIHT presented in the paper permits the decompression of only the needed image fragment. Herewith, the amount of memory used by the algorithm is comparable on order to the amount of memory used by the image fragment and not the image as a whole. Aside from this, these fragments can be extracted at various scales by the proposed method SS-SPIHT. Interactive analysis of huge images is possible due to this feature, i.e., viewing images the resolution of which in any dimension reaches several thousand or even hundred thousand points on a low-memory device, for instance, on a palm-held device.