Dual-Criterion Quality Loss for Blind Image Quality Assessment

Desen Yuan, Lei Wang

Abstract: This paper introduces a novel approach to Image Quality Assessment (IQA) by presenting a new loss function, Dual-Criterion Quality (DCQ) Loss, which integrates the Mean Squared Error (MSE) framework with a Relative Perception Constraint (RPC). The RPC is comprised of two main components: the Quantitative Discrepancy Constraint (QDC) and the Qualitative Alignment Constraint (QAC). The QDC focuses on capturing the numerical relationships of relative differences by minimizing the mean squared error between the differences in predicted scores among samples within a batch size and the differences in Mean Opinion Scores (MOS). Meanwhile, the QAC aims to capture the ordinal relationships between these differences. This method is designed to closely align with human subjective assessments of image quality, which are frequently quantified using the MOS, and to enhance the interpretability and reliability of IQA. Unlike existing ranking methods that suffer from complex pipelines and the introduction of errors through the generation of pair-wise or ordering data, DCQ Loss provides a more straightforward and efficient approach. Moreover, the loss function outperforms current rank-based IQA methods in terms of convergence, stability, and the ability to emulate human perception of visual quality. The effectiveness of this approach is validated through extensive experiments on various mainstream datasets and IQA network architectures, demonstrating significant performance gains over traditional rank loss approaches and contributing to the ongoing development of IQA.

Primary Subject Area: [Experience] Interactions and Quality of Experience

Secondary Subject Area: [Experience] Interactions and Quality of Experience

Relevance To Conference:

This work significantly advances the field of multimedia/multimodal processing, specifically within the scope of [Experience] Interactions and Quality of Experience, by addressing the critical area of Image Quality Assessment (IQA). Our research dives into the nuanced domain of assessing perceptual image quality, mirroring human subjective evaluations with high precision. By focusing on Blind IQA, our work facilitates quality assessment without reference images.
Our introduction of the novel Dual-Criterion Quality (DCQ) Loss stands at the forefront of this contribution, blending Mean Squared Error (MSE) with a unique Relative Perception Constraint to mimic human perceptual capabilities more closely. This methodology not only aligns with but also enriches the multimedia experience by ensuring that image quality assessments are both interpretative and reliable, resonating with the actual human perception of visual quality. This alignment is paramount in applications requiring high fidelity, such as medical imaging and video streaming services, where our approach significantly elevates the quality of experience.
By addressing the limitations of current ranking methods and introducing a more efficient, accurate, and human-centric approach, this work contributes profoundly to the multimedia/multimodal processing field, promising to enhance user experiences across a wide range of applications.

Submission Number: 422