Blind image quality assessment by simulating the visual cortex

Cai, Ruixian; Fang, Ming

doi:10.1007/s00371-022-02614-y

Cited by 9 publications

(1 citation statement)

References 58 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In today's rapidly evolving digital imaging landscape, IQA has emerged as a pivotal discipline at the intersection of technology and visual fidelity. Propelled by the swift advances in image acquisition and transmission technologies, coupled with an escalating demand for high-quality visual experiences, the evolution of IQA methods has gained considerable momentum [1][2][3][4]. Historically, IQA approaches have been rooted in emulating human visual perception, a strategy that has served well up to a point.…”

Section: Introductionmentioning

confidence: 99%

Innovative Approaches in Image Quality Assessment: A Deep Learning-Enabled Multi-Level and Multi-Scale Perspective

Sun

2023

View full text Add to dashboard Cite

In the dynamic field of digital image technology, the imperative role of Image Quality Assessment (IQA) is increasingly recognized. Traditional methodologies, designed to echo human visual processing, frequently encounter challenges in diverse application landscapes, primarily due to their singular focus on limited scale and level analysis. This shortcoming curtails their efficacy in practical scenarios. The incorporation of deep learning paradigms into IQA has notably enhanced evaluation capabilities. Yet, there remains a scope for refinement, especially in areas like integrating multi-scale data, fusing features at multiple levels, and optimizing computational resources. Addressing these gaps, this study proposes an advanced multi-level and multi-scale IQA strategy, harnessing the power of deep learning. A unique end-to-end multi-scale IQA module has been crafted, tailored to aggregate image quality data across a spectrum of scales comprehensively. Additionally, this research introduces an IQA model built upon the foundation of multi-level feature fusion. This innovative model stands out in its capacity to efficiently assess image quality, by adeptly extracting and amalgamating features from various levels. Beyond enhancing accuracy in quality scoring, this approach significantly bolsters the model's interpretability and operational efficiency, marking a stride forward in digital image processing research and applications.

show abstract

Section: Introductionmentioning

confidence: 99%