SPCM: Image quality assessment based on symmetry phase congruency

Abstract

Phase congruency (PC) algorithm is a frequency based algorithm. Instead of processing image spatially, the PC algorithm calculates the phase and amplitude of individual frequency components in the frequency domain. As one of the successful algorithm of image feature detection, PC has some advantages in the image quality assessment, however it has some inherent limitations. This paper studies the applications of symmetry phase in image quality assessment (IQA) and proposes a metric based on symmetry phase congruency (SPC). Symmetry phase congruency overcomes the limitations of phase congruency during the feature detection of image. This paper proposes a new IQA metric which is named as the symmetry phase congruency metric (SPCM). The sign responses of neighboring pixels are used to find the location of symmetry phases and then the symmetry phase congruency is used to detect image features and assess image quality. The experimental results show that SPCM is more sensitive to structural features of image and more robust to noises, and SPCM can achieve a higher consistency with the subjective evaluation of image quality.