Cross-Dataset-Robust Method for Blind Real-World Image Quality Assessment

TL;DR

This paper introduces a robust blind image quality assessment method that leverages a large-scale real-world dataset, a novel training strategy, and a powerful backbone to improve cross-dataset generalization.

Contribution

It proposes a new training approach using pseudo-labels from multiple models and a large dataset to enhance robustness and generalization in BIQA.

Findings

Outperforms some SOTA methods on cross-dataset tests

Uses a large dataset with 1,000,000 image pairs and pseudo-labels

Demonstrates improved robustness and generalization

Abstract

Although many effective models and real-world datasets have been presented for blind image quality assessment (BIQA), recent BIQA models usually tend to fit specific training set. Hence, it is still difficult to accurately and robustly measure the visual quality of an arbitrary real-world image. In this paper, a robust BIQA method, is designed based on three aspects, i.e., robust training strategy, large-scale real-world dataset, and powerful backbone. First, many individual models based on popular and state-of-the-art (SOTA) Swin-Transformer (SwinT) are trained on different real-world BIQA datasets respectively. Then, these biased SwinT-based models are jointly used to generate pseudo-labels, which adopts the probability of relative quality of two random images instead of fixed quality score. A large-scale real-world image dataset with 1,000,000 image pairs and pseudo-labels is then proposed for training the final cross-dataset-robust model. Experimental results on cross-dataset tests show that the performance of the proposed method is even better than some SOTA methods that are directly trained on these datasets, thus verifying the robustness and generalization of our method.