Source-free Domain Adaptation Framework Based on Confidence Constrained Mean Teacher for Fundus Image Segmentation DOI
Yanqin Zhang,

Ding Ma,

Xiangqian Wu

и другие.

Neurocomputing, Год журнала: 2024, Номер unknown, С. 129262 - 129262

Опубликована: Дек. 1, 2024

Язык: Английский

Cardiac cavity segmentation review in the past decade: Methods and future perspectives DOI
Feiyan Li, Weisheng Li, Yucheng Shu

и другие.

Neurocomputing, Год журнала: 2025, Номер 622, С. 129326 - 129326

Опубликована: Янв. 6, 2025

Язык: Английский

Процитировано

0
Semi-Supervised Multi-Task Learning for Interpretable Quality Assessment of Fundus Images DOI

Lucas Gabriel Telesco,

Danila Nejamkin,

Eloy Mata

и другие.

Опубликована: Янв. 1, 2025

Язык: Английский

Процитировано

0
A semi-supervised multi-task assisted method for ultrasound medical image segmentation DOI
Honghe Li, Jinzhu Yang,

Mingjun Qu

и другие.

Neurocomputing, Год журнала: 2025, Номер unknown, С. 130217 - 130217

Опубликована: Апрель 1, 2025

Язык: Английский

Процитировано

0
Learning AI-Driven Automated Blood Cell Anomaly Detection: Enhancing Diagnostics and Telehealth in Hematology DOI Creative Commons
Sami Naouali,

Oussama El Othmani

Journal of Imaging, Год журнала: 2025, Номер 11(5), С. 157 - 157

Опубликована: Май 16, 2025

Hematology plays a critical role in diagnosing and managing wide range of blood-related disorders. The manual interpretation blood smear images, however, is time-consuming highly dependent on expert availability. Moreover, it particularly challenging remote resource-limited settings. In this study, we present an AI-driven system for automated cell anomaly detection, combining computer vision machine learning models to support efficient diagnostics hematology telehealth contexts. Our architecture integrates segmentation (YOLOv11), classification (ResNet50), transfer learning, zero-shot identify categorize types abnormalities from images. Evaluated real annotated samples, the achieved high performance, with precision 0.98, recall 0.99, F1 score 0.98. These results highlight potential proposed enhance diagnostic capabilities clinical decision making underserved regions.

Язык: Английский

Процитировано

0
Deep learning based retinal vessel segmentation and hypertensive retinopathy quantification using heterogeneous features cross-attention neural network DOI Creative Commons
Xinghui Liu,

Hongwen Tan,

Wu Wang

и другие.

Frontiers in Medicine, Год журнала: 2024, Номер 11

Опубликована: Май 22, 2024

Retinal vessels play a pivotal role as biomarkers in the detection of retinal diseases, including hypertensive retinopathy. The manual identification these is both resource-intensive and time-consuming. fidelity vessel segmentation automated methods directly depends on fundus images' quality. In instances sub-optimal image quality, applying deep learning-based methodologies emerges more effective approach for precise segmentation. We propose heterogeneous neural network combining benefit local semantic information extraction convolutional long-range spatial features mining transformer structures. Such cross-attention structure boosts model's ability to tackle structures images. Experiments four publicly available datasets demonstrate our superior performance big potential retinopathy quantification.

Язык: Английский

Процитировано

2
Source-free Domain Adaptation Framework Based on Confidence Constrained Mean Teacher for Fundus Image Segmentation DOI
Yanqin Zhang,

Ding Ma,

Xiangqian Wu

и другие.

Neurocomputing, Год журнала: 2024, Номер unknown, С. 129262 - 129262

Опубликована: Дек. 1, 2024

Язык: Английский

Процитировано

0