Improving Medical Image Quality Using a Super-Resolution Technique with Attention Mechanism

Applied Sciences, Journal Year: 2025, Volume and Issue: 15(2), P. 867 - 867

Published: Jan. 17, 2025

Image quality plays a critical role in medical image analysis, significantly impacting diagnostic outcomes. Sharp and detailed images are essential for accurate diagnoses, but acquiring high-resolution often demands sophisticated costly equipment. To address this challenge, study proposes convolutional neural network (CNN)-based super-resolution architecture, utilizing melanoma dataset to enhance resolution through deep learning techniques. The proposed model incorporates self-attention block that combines channel spatial attention emphasize important features. Channel uses global average pooling fully connected layers high-frequency features within channels. Meanwhile, applies single-channel convolution the domain. By integrating various blocks, feature extraction is optimized further expanded subpixel produce high-quality images. L1 loss generate realistic smooth outputs, outperforming existing methods capturing contours textures. Evaluations with ISIC 2020 dataset—containing 33126 training 10982 test skin lesion analysis—showed 1–2% improvement peak signal-to-noise ratio (PSNR) compared very (VDSR) enhanced (EDSR) architectures.

Language: Английский

Potential of Proliferative Markers in Pancreatic Cancer Management: A Systematic Review

Health Science Reports, Journal Year: 2025, Volume and Issue: 8(3)

Published: March 1, 2025

Pancreatic cancer is an aggressive malignancy with poor prognosis and limited treatment options. Chemotherapy remains a primary therapeutic approach, but patient responses vary significantly, emphasizing the need for reliable biomarkers. This review explores potential role of proliferative markers, including Ki-67, PCNA, Cyclin D1, PHH3, as predictive prognostic indicators in pancreatic management, aiming to enhance personalized strategies. We conducted narrative by searching Scopus, PubMed, Google Scholar studies focusing on PHH3 relation chemotherapy. The literature was reviewed evaluate these markers predicting chemotherapy response, tumor progression, overall survival. highlights clinical significance markers. Ki-67 PCNA are associated cell proliferation, while D1 regulates cycle progression linked mitotic activity. High expression levels often correlate increased aggressiveness poorer outcomes. Moreover, they show promise which can inform tailored However, challenges remain, standardization detection methods determination optimal cutoff values. Proliferative such hold tools management. Their integration into practice could improve accuracy decisions Further research validation necessary overcome existing optimize their application oncology.

Language: Английский