Blind Image Quality Assessment With Active Inference

TLDR

Blind image quality assessment is valuable yet difficult, and recent work suggests that emulating the human visual system’s active inference of primary image content can improve BIQA. This paper proposes a novel BIQA metric that mimics the active inference process of the internal generative mechanism. The method uses a GAN‑based active inference module to predict primary content by optimizing semantic similarity and structural completeness, then a multi‑stream CNN evaluates scene, distortion, and degradation aspects to compute quality. The proposed metric achieves competitive performance on five IQA databases and significant improvements in cross‑database evaluations.

Abstract

Blind image quality assessment (BIQA) is a useful but challenging task. It is a promising idea to design BIQA methods by mimicking the working mechanism of human visual system (HVS). The internal generative mechanism (IGM) indicates that the HVS actively infers the primary content (i.e., meaningful information) of an image for better understanding. Inspired by that, this paper presents a novel BIQA metric by mimicking the active inference process of IGM. Firstly, an active inference module based on the generative adversarial network (GAN) is established to predict the primary content, in which the semantic similarity and the structural dissimilarity (i.e., semantic consistency and structural completeness) are both considered during the optimization. Then, the image quality is measured on the basis of its primary content. Generally, the image quality is highly related to three aspects, i.e., the scene information (content-dependency), the distortion type (distortion-dependency), and the content degradation (degradation-dependency). According to the correlation between the distorted image and its primary content, the three aspects are analyzed and calculated respectively with a multi-stream convolutional neural network (CNN) based quality evaluator. As a result, with the help of the primary content obtained from the active inference and the comprehensive quality degradation measurement from the multi-stream CNN, our method achieves competitive performance on five popular IQA databases. Especially in cross-database evaluations, our method achieves significant improvements.