Visibility and Distortion Measurement for No-Reference Dehazed Image Quality Assessment via Complex Contourlet Transform

Recently, most dehazed image quality assessment (DQA) methods mainly focus on the estimation of remaining haze, omitting impact distortions from side effect dehazing algorithms, which lead to their limited performance. Addressing this problem, we proposed a learning both Visibility and Distortion Aware features no-reference (NR) Dehazed Quality Assessment method (VDA-DQA). aware are exploited characterize clarity optimization after dehazing, including brightness, contrast, sharpness feature extracted by complex contourlet transform (CCT). Then, distortion employed measure artifacts images, normalized histogram local binary pattern (LBP) reconstructed statistics CCT sub-bands corresponding chroma saturation map. Finally, all above mapped into scores support vector regression (SVR). Extensive experimental results six public DQA datasets verify superiority VDA-DQA in terms consistency with subjective visual perception, outperforms state-of-the-art methods.The source code is available at https://github.com/li181119/VDA-DQA.