Towards Thousands to One Reference: Can We Trust the Reference Image for Quality Assessment?

Traditional full-reference image quality assessment (FR-IQA) methods predict the perceptual quality of a distorted image with a given pristine-quality image as the reference. However, the near-threshold visual perception suggests that there could be numerous pristine-quality representations that are indistinguishable in a scene, and the so-called pristine image used in FR-IQA for reference is just one of them. With numerous approaches proposed for FR-IQA by evaluating the perceptual similarity, much less work has been dedicated to locating the best reference for the deterministic perceptual similarity measure. This article aims to answer the question that whether enabling the freedom in reference image selection could lead to better performance by designing a new FR-IQA paradigm FLexible REference (FLRE). The FLRE paradigm is developed in the feature space by attempting to obtain the feature-level reference of the distorted image via the selection of its corresponding best explanation within an equal-quality space. To this end, we devise the Perceptually Near-Threshold Estimation (PNTE) and the Pseudo-Reference Search (PRS) strategies. In particular, the PNTE module predicts the equal-quality map of a given pristine-quality feature, forming an equal-quality space. Subsequently, the PRS strategy is employed to locate the reference of the distorted feature within the equal-quality space in an element-wise minimum distance search manner. Due to the lack of the ground-truth reference (i.e., best explanation) of each distorted image, we optimize the pseudo-reference feature learning under three constraints, i.e., the quality regression loss, the disturbance maximization loss, and the content loss. We implement the FLRE as a plug-in module before the deterministic FR-IQA process, and experimental results have demonstrated that combining FLRE with the existing deep feature-based FR-IQA models can significantly improve the quality prediction performance, largely surpassing the state-of-the-art methods.