Opinion-Unaware Blind Image Quality Assessment using Multi-Scale Deep Feature Statistics

TL;DR

This paper introduces a novel opinion-unaware blind image quality assessment method that combines deep features from pre-trained models with statistical analysis, eliminating the need for human ratings and enhancing efficiency and generalizability.

Contribution

It proposes a multi-scale deep feature statistics model that assesses image quality without human ratings by using pre-trained visual features and Gaussian modeling.

Findings

Outperforms state-of-the-art BIQA models in consistency with human perception

Demonstrates strong generalization across diverse BIQA datasets

Eliminates reliance on human rating data, improving training efficiency

Abstract

Deep learning-based methods have significantly influenced the blind image quality assessment (BIQA) field, however, these methods often require training using large amounts of human rating data. In contrast, traditional knowledge-based methods are cost-effective for training but face challenges in effectively extracting features aligned with human visual perception. To bridge these gaps, we propose integrating deep features from pre-trained visual models with a statistical analysis model into a Multi-scale Deep Feature Statistics (MDFS) model for achieving opinion-unaware BIQA (OU-BIQA), thereby eliminating the reliance on human rating data and significantly improving training efficiency. Specifically, we extract patch-wise multi-scale features from pre-trained vision models, which are subsequently fitted into a multivariate Gaussian (MVG) model. The final quality score is determined by quantifying the distance between the MVG model derived from the test image and the benchmark MVG model derived from the high-quality image set. A comprehensive series of experiments conducted on various datasets show that our proposed model exhibits superior consistency with human visual perception compared to state-of-the-art BIQA models. Furthermore, it shows improved generalizability across diverse target-specific BIQA tasks. Our code is available at: https://github.com/eezkni/MDFS