Perceptual objective evaluation for multimodal medical image fusion DOI Creative Commons

Chuangeng Tian,

Jun Zhang,

Lu Tang

и другие.

Frontiers in Physics, Год журнала: 2025, Номер 13

Опубликована: Май 26, 2025

Multimodal medical Image fusion (MMIF) has received widespread attention due to its promising application in clinical diagnostics and treatment. Due the inherent limitations of algorithms, quality obtained fused images (MFI) varies significantly. An objective evaluation MMIF can quantify visual differences facilitate rapid development advanced techniques, thereby enhancing image quality. However, rare research been dedicated evaluation. In this study, we present a multi-scale aware network for Specifically, employ Multi-scale Transform structure that simultaneously processes these using an ImageNet pre-trained ResNet34. Subsequently, incorporate online class activation mapping mechanism focus on lesion region, representative discrepancy features closely associated with MFI Finally, aggregate enhanced map them difference. lack dataset task, collect 129 pairs source from public datasets, namely, Whole Brain Atlas, construct database containing 1,290 generated algorithms. Each was annotated subjective score by experienced radiologists. Experimental results demonstrate our method produces satisfactory consistent perception, superior state-of-the-art methods. The is publicly available at: http://www.med.harvard.edu/AANLIB/home.html .

Язык: Английский

A Systematic Review of Medical Image Quality Assessment DOI Creative Commons

H. M. S. S. Herath,

H.M.K.K.M.B. Herath, Nuwan Madusanka

и другие.

Journal of Imaging, Год журнала: 2025, Номер 11(4), С. 100 - 100

Опубликована: Март 27, 2025

Medical image quality assessment (MIQA) is vital in medical imaging and directly affects diagnosis, patient treatment, general clinical results. Accurate high-quality necessary to make accurate diagnoses, efficiently design treatments, consistently monitor diseases. This review summarizes forty-two research studies on diverse MIQA approaches their effects performance diagnostics, results, efficiency the process. It contrasts subjective (manual assessment) objective (rule-driven) evaluation methods, underscores growing promise of machine intelligence learning (ML) automation, describes existing challenges. AI-powered tools are revolutionizing with automated checks, noise reduction, artifact removal, producing consistent reliable evaluation. Enhanced demonstrated every examination improve diagnostic precision support decision making clinic. However, challenges still exist, such as variability human ratings small datasets hindering standardization. These must be addressed better-quality data, low-cost labeling, Ultimately, this paper reinforces need for potential power AI. crucial advance area healthcare.

Язык: Английский

Процитировано

0
Perceptual objective evaluation for multimodal medical image fusion DOI Creative Commons

Chuangeng Tian,

Jun Zhang,

Lu Tang

и другие.

Frontiers in Physics, Год журнала: 2025, Номер 13

Опубликована: Май 26, 2025

Multimodal medical Image fusion (MMIF) has received widespread attention due to its promising application in clinical diagnostics and treatment. Due the inherent limitations of algorithms, quality obtained fused images (MFI) varies significantly. An objective evaluation MMIF can quantify visual differences facilitate rapid development advanced techniques, thereby enhancing image quality. However, rare research been dedicated evaluation. In this study, we present a multi-scale aware network for Specifically, employ Multi-scale Transform structure that simultaneously processes these using an ImageNet pre-trained ResNet34. Subsequently, incorporate online class activation mapping mechanism focus on lesion region, representative discrepancy features closely associated with MFI Finally, aggregate enhanced map them difference. lack dataset task, collect 129 pairs source from public datasets, namely, Whole Brain Atlas, construct database containing 1,290 generated algorithms. Each was annotated subjective score by experienced radiologists. Experimental results demonstrate our method produces satisfactory consistent perception, superior state-of-the-art methods. The is publicly available at: http://www.med.harvard.edu/AANLIB/home.html .

Язык: Английский

Процитировано

0