Enhanced Image Quality Assessment based on the Joint Similarity Feature

摘要

Most existing image quality assessment algorithms are designed for distorted images, while there is no special assessment for enhanced images. However, the existing assessment indices are related to image enhancement methods. In an attempt to design a quality assessment algorithm for enhanced images, deep research into the correlation between image enhancement and image quality has been carried out. Based on the four major features of image enhancement methods, which are lightness, contrast, saturation and sharpness, a new enhanced image quality assessment (EIQA) index which combines multiple similar features is proposed. Experimental results show that the proposed assessment index has a good consistency with the subjective score and has excellent performance for enhanced image quality assessment. In current research, the SROCC (Spearman's rank correlation coefficient) and PLCC (Pearson linear correlation coefficient) of the proposed index are both greater than 0.7. Moreover, the algorithm has high operating efficiency. (c) 2017 Society for Imaging Science and Technology.