Objective Quality Evaluation Method Of Stereo Image Based On HVS And Region Matching

Compared with 2-D images, a more immersive experience and telepresence can be given by stereo images. Recent years,the technology of stereo imaging has become a popular research topic at home and abroad. However, in a stereoscopic video and image acquisition, transmission, storage process, the amount of data at least doubled. How to ensure the stereoscopic video and image quality will directly affect the development of stereo imaging technology. Therefore, it is significant to propose an accurate and efficient objective assessment method for stereo images.This evaluation method or standard not only indicates damage degree of the processing and transmission technology to the stereo images, but also describes accuracy of stereo display technology.From the plane of the image quality assessment research present situation and the development direction analysis, this paper deeply studied human visual characteristics and stereo imaging theory, established a human visual system model which reflected luminance sensitivity, contrast sensitivity, multi-channel effects and masking effects. At last, the objective evaluation indications of stereo image quality were abstracted,including Steerable Pyramid, Structural Similarity and the regional feature point matching ratio used as reflecting the stereo sense. We proposed a reasonable total evaluation metric by appropriate weighted method.The result showed that our method evaluating the stereo image quality of the common type of distortion had a high correlation with human experimental score, which could reflect stereo image quality correctly, and overcome shortcomings of subjective evaluation method which was operation complexly, time-consuming and application poorly. This method has referenced value and practical significance to promote the development of evaluation method of stereo images quality.