Neurocomputing, Год журнала: 2023, Номер 543, С. 126285 - 126285
Опубликована: Апрель 27, 2023
Язык: Английский
Neurocomputing, Год журнала: 2023, Номер 543, С. 126285 - 126285
Опубликована: Апрель 27, 2023
Язык: Английский
Computers in Biology and Medicine, Год журнала: 2022, Номер 151, С. 106265 - 106265
Опубликована: Ноя. 9, 2022
Язык: Английский
Процитировано
44Expert Systems with Applications, Год журнала: 2023, Номер 237, С. 121305 - 121305
Опубликована: Сен. 4, 2023
Язык: Английский
Процитировано
40Expert Systems with Applications, Год журнала: 2023, Номер 229, С. 120477 - 120477
Опубликована: Май 17, 2023
In December 2019, the global pandemic COVID-19 in Wuhan, China, affected human life and worldwide economy. Therefore, an efficient diagnostic system is required to control its spread. However, automatic poses challenges with a limited amount of labeled data, minor contrast variation, high structural similarity between infection background. this regard, new two-phase deep convolutional neural network (CNN) based proposed detect minute irregularities analyze infection. first phase, novel SB-STM-BRNet CNN developed, incorporating channel Squeezed Boosted (SB) dilated convolutional-based Split-Transform-Merge (STM) block infected lung CT images. The STM blocks performed multi-path region-smoothing boundary operations, which helped learn variation specific patterns. Furthermore, diverse boosted channels are achieved using SB Transfer Learning concepts texture COVID-19-specific healthy second images provided COVID-CB-RESeg segmentation identify infectious regions. methodically employed region-homogeneity heterogeneity operations each encoder-decoder boosted-decoder auxiliary simultaneously low illumination boundaries region. yields good performance terms accuracy: 98.21 %, F-score: 98.24%, Dice Similarity: 96.40 IOU: 98.85 % for would reduce burden strengthen radiologist's decision fast accurate diagnosis.
Язык: Английский
Процитировано
32Neurocomputing, Год журнала: 2023, Номер 548, С. 126411 - 126411
Опубликована: Июнь 2, 2023
Язык: Английский
Процитировано
24IEEE/CAA Journal of Automatica Sinica, Год журнала: 2024, Номер 11(6), С. 1536 - 1538
Опубликована: Май 27, 2024
Dear Editor, In this letter, a novel hierarchical fusion framework is proposed to address the imperfect data property in complex medical image analysis (MIA) scenes. particular, by combining strengths of convolutional neural networks (CNNs) and transformers, enhanced feature extraction, spatial modeling, sequential context learning are realized provide comprehensive insights on patterns. Integration information different level enabled via multi-attention mechanism, tensor decomposition methods adopted so that compact distinctive representation underlying high-dimensional features can be accomplished [1]. It shown from evaluation results competitive superior as compared with some other advanced algorithms, which effectively handles inter-class similarity intra-class differences diseases, meanwhile, model complexity reduced within an acceptable level, benefits deployment clinic practice.
Язык: Английский
Процитировано
12Neurocomputing, Год журнала: 2024, Номер 576, С. 127325 - 127325
Опубликована: Янв. 23, 2024
Язык: Английский
Процитировано
11International Journal of Network Dynamics and Intelligence, Год журнала: 2024, Номер unknown, С. 100009 - 100009
Опубликована: Июнь 26, 2024
Article UNet and Variants for Medical Image Segmentation Walid Ehab, Lina Huang, Yongmin Li * Department of Computer Science, Brunel University London, Uxbridge, UB8 3PH, United Kingdom Correspondence: [email protected] Received: 22 September 2023 Accepted: 25 December Published: 26 June 2024 Abstract: imaging plays a crucial role in modern healthcare by providing non-invasive visualisation internal structures abnormalities, enabling early disease detection, accurate diagnosis, treatment planning. This study aims to explore the application deep learning models, particularly focusing on architecture its variants, medical image segmentation. We seek evaluate performance these models across various challenging segmentation tasks, addressing issues such as normalization, resizing, choices, loss function design, hyperparameter tuning. The findings reveal that standard UNet, when extended with network layer, is proficient model, while Res-UNet Attention architectures demonstrate smoother convergence superior performance, handling fine details. also addresses challenge high class imbalance through careful preprocessing definitions. anticipate results this will provide useful insights researchers seeking apply new problems offer guidance best practices their implementation.
Язык: Английский
Процитировано
10Neurocomputing, Год журнала: 2022, Номер 513, С. 94 - 103
Опубликована: Сен. 21, 2022
Язык: Английский
Процитировано
39Engineering Applications of Artificial Intelligence, Год журнала: 2022, Номер 117, С. 105632 - 105632
Опубликована: Ноя. 21, 2022
Язык: Английский
Процитировано
37Neurocomputing, Год журнала: 2023, Номер 525, С. 57 - 75
Опубликована: Янв. 12, 2023
Язык: Английский
Процитировано
23