Article
Computer Science, Artificial Intelligence
Yibo Gao, Huan Wang, Zuhao Liu
Summary: In this study, a residual-based temporal attention convolutional neural network (RTA-CNN) was proposed for atrial fibrillation (AF) detection, which automatically focuses on parts with more semantic information to achieve better performance. Additionally, a novel exponential nonlinearity loss (EN-Loss) was introduced to address the imbalance problem.
KNOWLEDGE-BASED SYSTEMS
(2021)
Article
Automation & Control Systems
Meng Xiao, Bo Yang, Shilong Wang, Zhengping Zhang, Yan He
Summary: This paper proposes a Fine Coordinate Attention (FCA) block to address the challenges of surface defect detection. The FCA block encodes both average and salient information in two coordinate directions, capturing spatial dependence and achieving long-range interaction. Experimental results show that the FCA block outperforms existing attention mechanisms in image classification and object detection tasks.
ENGINEERING APPLICATIONS OF ARTIFICIAL INTELLIGENCE
(2023)
Article
Computer Science, Information Systems
Jiana Meng, Yanlin Zhu, Shichang Sun, Dandan Zhao
Summary: Sarcasm detection is a challenging task in sentiment analysis. To address the issue of complex semantic information and lack of sentiment words in sentences, we propose a sarcasm detection model based on pretraining model and attention mechanism, which extracts semantic features of phrase fragments based on context and language environment to improve the model's ability to identify sarcasm expressions.
MULTIMEDIA TOOLS AND APPLICATIONS
(2023)
Article
Biology
Amir Ebrahimi, Suhuai Luo, Raymond Chiong
Summary: The study examined the effectiveness of applying deep sequence-based network models for AD detection, addressing the classification accuracy issue of 2D and 3D CNNs in AD detection by handling the MRI feature sequences generated by CNNs.
COMPUTERS IN BIOLOGY AND MEDICINE
(2021)
Article
Automation & Control Systems
Dahang Wan, Rongsheng Lu, Siyuan Shen, Ting Xu, Xianli Lang, Zhijie Ren
Summary: This paper proposes a lightweight Mixed Local Channel Attention (MLCA) module to improve the performance of the object detection network. It can simultaneously incorporate both channel information and spatial information, as well as local information and global information to improve the expression effect of the network. The MobileNet-Attention-YOLO (MAY) algorithm is presented to compare the performance of various attention modules. MLCA achieves a better balance between model representation efficacy, performance, and complexity than alternative attention techniques on the Pascal VOC and SMID datasets.
ENGINEERING APPLICATIONS OF ARTIFICIAL INTELLIGENCE
(2023)
Article
Engineering, Marine
Yuchao Wang, Jingdong Li, Zeming Chen, Chenglong Wang
Summary: The YOLOX algorithm combined with Convolutional Block Attention Module (CBAM) is proposed to improve the low accuracy of small target detection in traditional algorithms. Experimental results show that the proposed CBAM-YOLOX network improves the average accuracy and recall rate of target detection, validating its effectiveness.
JOURNAL OF MARINE SCIENCE AND ENGINEERING
(2022)
Article
Agricultural Engineering
WenXin Hu, JunTao Xiong, JunHao Liang, ZhiMing Xie, ZhiYu Liu, QiYin Huang, ZhenGang Yang
Summary: Achieving accurate detection of postharvest citrus epidermal defects is crucial for ensuring the quality and value of citrus. The uneven lighting conditions pose a challenge to detecting citrus epidermis abnormalities. In this study, a dual-lamp image acquisition system was designed for capturing images of citrus fruit invisible defects, and the YOLOv5 model was optimized to improve the detection accuracy. Experimental results showed that the improved YOLOv5 model outperformed YOLOv5x, achieving higher mAP, Precision, and Recall values, as well as an increased average detection speed. This research provides technical support for intelligent detection and grading of postharvest citrus.
BIOSYSTEMS ENGINEERING
(2023)
Article
Computer Science, Artificial Intelligence
Yingying Zhu, Yinghao Wang, Haonan Chen, Zemian Guo, Qiang Huang
Summary: This paper proposes a method based on attention mechanism to address the problem of clutter and occlusion in feature extraction when using convolutional neural networks (CNN). Two attention modules, spatial attention module and channel attention module, are introduced to adjust the weight distribution of feature maps, making the extracted features more discriminative. Furthermore, a scale and mask scheme is presented to filter out redundant features and reduce the disadvantages of various scales. Experimental results demonstrate the effectiveness of the proposed method on four well-known image retrieval datasets.
INTERNATIONAL JOURNAL OF NEURAL SYSTEMS
(2023)
Article
Neurosciences
Hao Guan, Chaoyue Wang, Jian Cheng, Jing Jing, Tao Liu
Summary: The proposed deep learning framework for AD diagnosis learns global and local features directly from sMRI scans without prior knowledge, achieving competitive results on public datasets. The framework is lightweight, suitable for end-to-end training, and effective when medical priors are unavailable or computing resources are limited.
HUMAN BRAIN MAPPING
(2022)
Article
Computer Science, Artificial Intelligence
Jin Zhang, Yanjiao Shi, Qing Zhang, Liu Cui, Ying Chen, Yugen Yi
Summary: The article introduces a new neural network called ACFFNet for salient object detection, which includes MCA module, CFF module, and SR module, as well as a CCE loss to guide the network to focus on more detailed information. Experimental results show that the proposed method outperforms the state-of-the-art methods.
IMAGE AND VISION COMPUTING
(2022)
Article
Environmental Sciences
Jinming Wu, Chunhui Xie, Zuxi Zhang, Yongxin Zhu
Summary: This paper proposes a deeply supervised attentive high-resolution network (DSAHRNet) for remote sensing image change detection. It decodes change information from bitemporal features using a spatial-channel attention module, reduces feature misalignment through stacked convolutional blocks, and generates more discriminative features using a novel deeply supervised module. Experimental results on three challenging benchmark datasets demonstrate that DSAHRNet outperforms other state-of-the-art methods and achieves a great trade-off between performance and complexity.
Article
Agriculture, Multidisciplinary
Qixin Sun, Xiujuan Chai, Zhikang Zeng, Guomin Zhou, Tan Sun
Summary: This paper proposes a noise-tolerant feature fusion network (NTFFN) for outdoor fruit detection and constructs an RGB-D citrus fruit dataset. Evaluation results show that NT-FFN outperforms other methods in terms of performance, validating its generalization ability in fruit detection tasks.
COMPUTERS AND ELECTRONICS IN AGRICULTURE
(2022)
Article
Chemistry, Multidisciplinary
Feiyun Wang, Chengxu Lv, Yuxuan Pan, Liming Zhou, Bo Zhao
Summary: This study proposed the combination of ResNet and CBAM for the automatic detection of external defects in kiwifruit. The results showed that ResNet34 + CBAM achieved high recognition accuracy and good stability for kiwifruit defect detection.
APPLIED SCIENCES-BASEL
(2023)
Article
Computer Science, Artificial Intelligence
Lei Li, Juan Qin, Lianrong Lv, Mengdan Cheng, Biao Wang, Dan Xia, Shike Wang
Summary: In recent years, there has been increasing attention on the spine and related diseases. Spinal parsing plays an important role in the diagnosis and treatment of various spinal diseases. This paper proposes an efficient and novel automatic segmentation network model for MR spine images. The model utilizes an Inception structure and attention mechanisms to improve the segmentation performance. Experimental results show significant improvements in the segmentation indicators, indicating the effectiveness of the model.
INTERNATIONAL JOURNAL OF MACHINE LEARNING AND CYBERNETICS
(2023)
Article
Computer Science, Artificial Intelligence
Lu Zhang, Fengling Jiang, Bin Kong, Jing Yang, Can Wang
Summary: This paper proposes a real-time lane detection method using attention mechanism, and comparative experiments on two benchmark datasets show that the method is independent of lane number and pattern, running at high speed with superior performance.
COGNITIVE COMPUTATION
(2021)
Article
Radiology, Nuclear Medicine & Medical Imaging
Slimane Tounekti, Mahdi Alizadeh, Devon Middleton, James S. Harrop, Bassem Hiba, Laura Krisa, Choukri Mekkaoui, Feroze B. Mohamed
Summary: This study proposes and demonstrates a new method combining reduced field-of-view strategy with phase segmented EPI to address geometric distortion in post-operative DTI scans of patients with metal implants. The results show that the new method outperforms traditional techniques in reducing distortion.
MAGNETIC RESONANCE IMAGING
(2024)
Article
Radiology, Nuclear Medicine & Medical Imaging
Silvia Minosse, Eliseo Picchi, Valentina Ferrazzoli, Noemi Pucci, Valerio Da Ros, Raffaella Giocondo, Roberto Floris, Francesco Garaci, Francesca Di Giuliano
Summary: The aim of this study was to investigate the variation of DCE-MRI-derived kinetic parameters in brain tumors as a function of acquisition time. The results showed that K-ep and V-e were time-dependent and required longer scan times to obtain reliable parameter values, while K-trans was time-independent and remained the same in all acquisition times, making it a reliable parameter for short acquisition times.
MAGNETIC RESONANCE IMAGING
(2024)
Article
Radiology, Nuclear Medicine & Medical Imaging
Xingmin Guan, Xinheng Zhang, Hsin-Jung Yang, Rohan Dharmakumar
Summary: This study aims to investigate why DIR-prepared dark-blood T2* weighted images have lower SNR, CNR, and diagnostic accuracy for intramyocardial hemorrhage (IMH) detection compared to non-DIR-prepared bright-blood T2* images. Through phantom and animal studies, it was confirmed that the signal loss on DIR-prepared T2* images mainly originates from spin-relaxation during the DIR preparation. Therefore, when used for IMH detection, extra attention should be paid to the SNR of DIR-prepared dark-blood T2* imaging protocols.
MAGNETIC RESONANCE IMAGING
(2024)
Article
Radiology, Nuclear Medicine & Medical Imaging
Beatriz Laureano, Hassna Irzan, Helen OReilly, Sebastian Ourselin, Neil Marlow, Andrew Melbourne
Summary: Prematurity and preterm stressors have significant effects on the development of infants, especially at earlier gestations. While neonatal care advances have reduced preterm mortality rates, disability rates continue to grow in middle-income settings. Imaging the preterm brain using MR technology has improved our understanding of its development and the affected regions and networks. This research aims to support interventions, improve neurodevelopment, and provide accurate prognoses for preterm infants. This study focuses on the fully developed brain of extremely preterm subjects and examines myelin-related biomarkers to assess long-term effects. The findings suggest altered connectivity and cognitive outcomes in the adult preterm brain.
MAGNETIC RESONANCE IMAGING
(2024)
Article
Radiology, Nuclear Medicine & Medical Imaging
Julian Rauch, Frederik B. Laun, Peter Bachert, Mark E. Ladd, Tristan A. Kuder
Summary: This study presents a method for reducing concomitant field effects in double diffusion encoding (DDE) sequences by adding oscillating gradient pulses. The modified sequences successfully reduced accumulated concomitant phase without significant changes in the original sequence characteristics. The proposed method led to an increase in signal-to-noise ratio (SNR) for phantom and in vivo experiments, supported by simulations.
MAGNETIC RESONANCE IMAGING
(2024)
Article
Radiology, Nuclear Medicine & Medical Imaging
Marlon Bran Lorenzana, Shekhar S. Chandra, Feng Liu
Summary: Sparse reconstruction is important in MRI for reducing acquisition time and improving spatial-temporal resolution. This paper introduces two decoupling techniques for explicit 1D regularization and a combined 1D + 2D reconstruction technique that improves image quality.
MAGNETIC RESONANCE IMAGING
(2024)
Article
Radiology, Nuclear Medicine & Medical Imaging
Yifan Gou, W. Christopher Golden, Zixuan Lin, Jennifer Shepard, Aylin Tekes, Zhiyi Hu, Xin Li, Kumiko Oishi, Marilyn Albert, Hanzhang Lu, Peiying Liu, Dengrong Jiang
Summary: ARTS algorithm improves the reliability of Y-v estimation in noncompliant subjects, enhancing the utility of Y-v as a biomarker for brain diseases.
MAGNETIC RESONANCE IMAGING
(2024)