Variable resolution-adaptive sampling approach to image coding

Since the local characteristics of an image usually change from region to region, adaptive coding techniques are necessary to achieve high compression rates. An effective coding scheme is based on the principle of variable resolution image representation, where compression is generally achieved by adaptive sampling or subsampling. Various adaptive sampling schemes that are effective for image compression are developed and evaluated. The main focus is the design of the nonuniform sampling patterns associated with the sample selection process. Two different design procedures, based on nonrecursive and recursive image decomposition, are investigated. Simulation results indicate that some of the sampling schemes proposed here attain higher compression rates than other reported coders of this nature