Fast Lifting IQA Algorithm of the Structure Similarity Index

Image quality assessment（IQA） aims to design mathematical models for measuring an image quality well consistent with subjective evaluations. Considering both the limitations of structure similarity（SSIM） index and other methods, and the advantages of the gradient on characterizing the edge and texture information of an image, we put forward a kind of fast full-reference IQA algorithm, namely gradient weighted lifting SSIM（GWL-SSIM） method. Firstly, we define the generalized gradient of local image block. Then employ the generalized gradient similarity, contrast similarity and structural similarity to obtain a feature mapping image of local quality. Finally, we adopt the weighted pooling strategy by generalized gradient information to obtain an image evaluation model. In our algorithm, the generalized gradient well describes the nonlinearity of human visual system（HVS） for image perception, while the weighted strategy simulates the diversity how HVS perceive the different ingredients of an image. Numerical experiments, performed on six public databases, demonstrate that GWL-SSIM can achieve high computational efficiency and have a considerable evaluation result compared with state-of-the-art IQA metrics.