Approach for enhancing the accuracy of semantic segmentation of chest X-ray images by edge detection and deep learning integration DOI Creative Commons
Lesia Mochurad

Frontiers in Artificial Intelligence, Journal Year: 2025, Volume and Issue: 8

Published: April 16, 2025

Introduction Accurate segmentation of anatomical structures in chest X-ray images remains challenging, especially for regions with low contrast and overlapping structures. This limitation significantly affects the diagnosis cardiothoracic diseases. Existing deep learning methods often struggle preserving structural boundaries, leading to artifacts. Methods To address these challenges, I propose a novel approach that integrates contour detection techniques U-net architecture. Specifically, method employs Sobel Scharr edge filters enhance boundaries before segmentation. The pipeline involves pre-processing using detection, followed by model trained identify lungs, heart, clavicles. Results Experimental evaluation demonstrated edge-enhancing filters, particularly operator, leads marked improvement accuracy. For lung segmentation, achieved an accuracy 99.26%, Dice coefficient 98.88%, Jaccard index 97.54%. Heart results included 99.47% 94.14% index, while clavicle reached 99.79% 89.57% index. These consistently outperform baseline without enhancement. Discussion integration improves quality complex X-rays. Among tested operator proved be most effective enhancing boundary information reducing offers promising direction more accurate robust computer-aided systems radiology.

Language: Английский

Variability in Arterial Stiffness and Vascular Endothelial Function After COVID-19 During 1.5 Years of Follow-Up—Systematic Review and Meta-Analysis DOI Creative Commons
Danuta Łoboda, Krzysztof S. Gołba, Piotr Jerzy Gurowiec

et al.

Life, Journal Year: 2025, Volume and Issue: 15(4), P. 520 - 520

Published: March 21, 2025

Increasing long-term observations suggest that coronavirus disease 2019 (COVID-19) vasculopathy may persist even 1.5 years after the acute phase, potentially accelerating development of atherosclerotic cardiovascular diseases. This study systematically reviewed variability brachial flow-mediated dilation (FMD) and carotid-femoral pulse wave velocity (cfPWV) from phase COVID-19 through 16 months follow-up (F/U). Databases including PubMed, Web Science, MEDLINE, Embase were screened for a meta-analysis without language or date restrictions (PROSPERO reference CRD42025642888, last search conducted on 1 February 2025). The quality included studies was assessed using Newcastle–Ottawa Quality Scale. We considered all (interventional pre-post studies, prospective observational randomized, non-randomized trials) FMD cfPWV in adults (aged ≥ 18 years) with laboratory-confirmed compared non-COVID-19 controls changes these parameters during F/U. Twenty-one reported differences FMD, examined between patients control groups various stages: acute/subacute (≤30 days onset), early (>30–90 days), mid-term (>90–180 late (>180–270 very (>270 days) post-COVID-19 recovery. Six while nine did so Data 14 (627 cases 694 controls) 15 (578 703 our meta-analysis. showed significant decrease to (standardized mean difference [SMD]= −2.02, p < 0.001), partial improvements noted recovery (SMD = 0.95, 0.001) 0.92, 0.006). Normalization observed 0.12, 0.69). In contrast, values, which higher than 1.27, remained elevated throughout F/U, no except (SMD= −0.39, 0.001). recovery, values those 0.45, 0.010). manuscript, we discuss how factors, severity COVID-19, persistence syndrome, patient’s initial vascular age, depending metrics age risk influenced time degree improvement.

Language: Английский

Citations

0
Approach for enhancing the accuracy of semantic segmentation of chest X-ray images by edge detection and deep learning integration DOI Creative Commons
Lesia Mochurad

Frontiers in Artificial Intelligence, Journal Year: 2025, Volume and Issue: 8

Published: April 16, 2025

Introduction Accurate segmentation of anatomical structures in chest X-ray images remains challenging, especially for regions with low contrast and overlapping structures. This limitation significantly affects the diagnosis cardiothoracic diseases. Existing deep learning methods often struggle preserving structural boundaries, leading to artifacts. Methods To address these challenges, I propose a novel approach that integrates contour detection techniques U-net architecture. Specifically, method employs Sobel Scharr edge filters enhance boundaries before segmentation. The pipeline involves pre-processing using detection, followed by model trained identify lungs, heart, clavicles. Results Experimental evaluation demonstrated edge-enhancing filters, particularly operator, leads marked improvement accuracy. For lung segmentation, achieved an accuracy 99.26%, Dice coefficient 98.88%, Jaccard index 97.54%. Heart results included 99.47% 94.14% index, while clavicle reached 99.79% 89.57% index. These consistently outperform baseline without enhancement. Discussion integration improves quality complex X-rays. Among tested operator proved be most effective enhancing boundary information reducing offers promising direction more accurate robust computer-aided systems radiology.

Language: Английский

Citations

0