质量评价（英文版）.pdfVIP

An SVD-Based Gray-Scale Image Quality Measure for Local and Global Assessment Aleksandr Shnayderman, Alexander Gusev, Ahmet M. Eskicioglu measures for image compression [14,16,28,31,40,44], and for Abstract—The important criteria used in subjective evaluation noise and blur [13,45]. The most common objective of distorted images include the amount of distortion, the type of evaluation tool, the MSE, is very unreliable, resulting in poor distortion, and the distribution of error. An ideal image quality correlation with the HVS. In spite of their complicated measure should therefore be able to mimic the human observer. algorithms, the HVS-based objective measures do not appear We present a new gray-scale image quality measure that can be to be superior to the simple pixel-based measures like the used as a graphical or a scalar measure to predict the distortion MSE, Peak Signal-to-Noise Ratio (PSNR), or Root Mean introduced by a wide range of noise sources. Based on the Singular Value Decomposition, it reliably measures the distortion Squared Error (RMSE). not only within a distortion type at different distortion levels but A number of researchers point to the disadvantages of the also across different distortion types. The measure was applied measures that incorporate an HVS model. Fuhrmann et al to five test images (Airplane, Boat, Goldhill, Lena, and Peppers) [44] discourage the use of metrics based on the spati