The Study Of Image Quality Assessment And It's Application

The study on image quality assessment is important in theory and widely required in applications, due to that image is subject to a variety of distortions during gathering, transmission, processing and recording.The thesis introduces the subjective and objective methods of the image quality assessment and analyzes a structural similarity (SSIM) metric in detail, firstly. Then based on the study of the relationship between multi-channels of human visual system and wavelet transform, an improved algorithm of SSIM named weight wavelet structural similarity (WW-SSIM) is proposed. The algorithm calculates the similarity of all sub-bands of the image after wavelet transformation. According to the different visual sensitivities of human beings to different frequency and orientation, the method carries the visual weights on different sub-bands'similarity. Finally, the WW-SSIM metric is formulated. The experimental results illustrate that the proposed metric approximates the mean of score (MOS) more than the metric of the SSIM.In the application of the image denoising system, a quality assessment metric can be used to guide the adjustment of the image processing system or to optimize the parameters of the algorithms. Therefore, it is of significance to design and develop the software of image quality assessment. In this thesis, the software module has been achieved and successfully applied to estimating the performance of the image processing system. Finally, the thesis does some experiments on the remote sensing image based on the SSIM. The experimental results illustrate that the SSIM is suitable for the remote sensing image. To calculate the value of the SSIM conveniently, the thesis achieves the software interface of SSIM.