Study On The Algorithms For HVS-based Color Image Measurement

With the development of the multimedia and network technology, the techniques of image processing have been fully developed in recent years, especially for image compression. On the other hand, image quality measures haven't been improved correspondingly. The most popular image quality assessment (IQA) is still traditional methods whose representative is PSNR (Peak Signal-to-Noise Ratio). This kind of methods evaluate image quality by calculating the differences of distorted and reference image pixels. The cause of formation, attribute and distribution of error signal are taken into no account, so the two pictures which have the same PSNR may be very different in human eyes. Although the subjective IQA keeps good consistency with human vision, it is time- consuming, laborious. Besides, its result may be differs greatly because of observers'various preference. So the subjective IQA can not be used widely. Today image encoding pursues low error rate and high compression ratio, which needs effective IQA. Traditional IQAs became an obstacle of image processing's development.Aims to the problem mentioned above, this paper proposed an IQA for color images which is on the base of widely referring to the material about IQAs from home and abroad. The creative points of this method are expatiated in the following part.1,The algorithm is established in color IQA. Previous IQAs are mainly designed for gray level images or binary images. When it comes to color IQA, the usual solution is applying IQAs into three frames (RGB,YUV,etc) independently and adopting their average, which lead to an inaccurate result. In this paper, we choose a new model of color image to value images'definition; besides, the algorithm assess image quality based on the color information. Thus, the result inclines to the...