Image representation and compression with steered Hermite transforms

With limited effort, one can obtain valuable information on the subjective performance of image compression schemes. The choice of an appropriate experimental technique, however, is not always trivia!, since both the subjeet's task and the experimental metbod itself should depend on the properties of the stimulus set ( quality range, kind of artifacts). The goal of this chapter is to discuss a number of numerical sealing techniques that can be used to assess perceived image quality. It is shown how to analyze data obtained in numerical category sealing experirnents and how to set up such experirnents. The results of several subjective experiments illustrate that numerical category sealing techniques provide an effi.cient rneans not only for obtaining cornpression ratio versus quality curves that characterize coder performance over a braad range of compression ratios, but also for perceived image quality in a much smaller range (e.g., close to threshold level). However, the nature of the artifacts introduced by different coders can cause problems when evaluations are carried out using direct numerical category sealing. The latter is demonstrated by camparing the results of a direct sealing methad and a sealing technique in which subjects have to determine quality differences between all possible combinations of coded images. 1 The research described in this chapter has been published as a paper in Signa] Processing [36]. Some results were also presented at the International Conference for Image Processing ICIP-96, Lausanne [34] 14 Chapter 2 Subjective Quality Evaluation of Compressed Images