Article
Microbiology
Houwu Gong, Xiong You, Min Jin, Yajie Meng, Hanxue Zhang, Shuaishuai Yang, Junlin Xu
Summary: This paper studied the prediction of microbe-disease associations using a multi-data biological network and a graph neural network algorithm. The proposed Microbe-Disease Heterogeneous Network and GCNN4Micro-Dis model showed good predictive power and outperformed other methods in predicting microbe-disease associations. Case study on breast cancer further verified the accuracy of the model.
FRONTIERS IN MICROBIOLOGY
(2022)
Article
Computer Science, Artificial Intelligence
Mighty Abra Ayidzoe, Yongbin Yu, Patrick Kwabena Mensah, Jingye Cai, Kwabena Adu, Yifan Tang
Summary: Capsule network is important for image recognition, but the basic model lacks adaptability and struggles to extract important features from complex images, a problem that can be addressed by a novel variant of capsule network.
MACHINE VISION AND APPLICATIONS
(2021)
Article
Biology
V. Jahmunah, E. Y. K. Ng, Tan Ru San, U. Rajendra Acharya
Summary: In this study, an automated system was developed using convolutional neural network (CNN) and unique GaborCNN models for categorizing ECG signals into normal, CAD, MI, and CHF classes. High classification accuracies exceeding 98.5% were achieved by the CNN and GaborCNN models, with GaborCNN being preferred for its performance and reduced complexity. This study is the first to propose using GaborCNN for automated categorization of ECG signals into normal, CAD, MI, and CHF classes.
COMPUTERS IN BIOLOGY AND MEDICINE
(2021)
Article
Computer Science, Information Systems
Salinna Abdullah, Majid Zamani, Andreas Demosthenous
Summary: This paper presents a CNN-based speech enhancement algorithm with an adaptive filter design (CNN-AFD) using Gabor function and region-aware convolution. The proposed algorithm incorporates fixed Gabor functions into convolutional filters to model human auditory processing for improved denoising performance. It also explores skip convolution and activation analysis-wise pruning to reduce the high cost of inference of the CNN. The results show that the proposed CNN-AFD outperforms other baseline algorithms in terms of speech quality and intelligibility.
Article
Computer Science, Artificial Intelligence
Wanxiang Li, Zhiwu Shang, Shiqi Qian, Baoren Zhang, Jie Zhang, Maosheng Gao
Summary: The paper proposes a novel diagnosis method based on signal-to-image mapping and deep Gabor convolutional adaptive pooling network, improving the feature extraction and model generalization in vibration signal fault diagnosis.
EXPERT SYSTEMS WITH APPLICATIONS
(2022)
Article
Biology
Cheng Chen, Kangneng Zhou, Siyu Qi, Tong Lu, Ruoxiu Xiao
Summary: Vessel segmentation is important for vascular disease characterization and has received significant attention from researchers. We propose a Gabor ConvNet, integrating Gabor convolution kernels into a CNN architecture, for vessel segmentation. Our method outperforms advanced models with scores of 85.06%, 70.52%, and 67.11% on three vessel datasets, respectively. Ablation studies demonstrate that Gabor kernels have better vessel extraction ability than regular convolution kernels.
COMPUTERS IN BIOLOGY AND MEDICINE
(2023)
Article
Multidisciplinary Sciences
Afifa Khaled, Jian-Jun Han, Taher A. Ghaleb, Radman Mohamed
Summary: This paper proposes an improved method for brain segmentation using a 3D convolutional neural network model. The method includes multi-instance loss and Gabor filter banks. Evaluation results show that the model is more accurate and efficient in segmenting brain images than existing models.
ARABIAN JOURNAL FOR SCIENCE AND ENGINEERING
(2023)
Article
Engineering, Civil
M. T. Vu, A. Jardani
Summary: The paper introduces a new method using convolutional neural networks to map fracture network structure in a heterogeneous aquifer by inverting hydraulic head data. The developed neural network excels in handling highly nonlinear inverse functions and accurately maps different complexity levels of fractures embedded in a matrix with heterogeneous transmissivity. The accuracy of the inversion results depends on the synthetic dataset used for training the network, with stable interpretations even in the presence of data noise.
JOURNAL OF HYDROLOGY
(2022)
Article
Engineering, Biomedical
Yanan Xu, Yingle Fan
Summary: This research proposes a novel method for retinal blood vessel segmentation based on a dual-channel asymmetric convolutional neural network (CNN). The method combines thick and thin vessel extraction modules and fuses the segmentation results of thick and thin vessels. In experiments, the method achieves good performance on the DRIVE and CHASE_DB1 databases.
BIOCYBERNETICS AND BIOMEDICAL ENGINEERING
(2022)
Article
Computer Science, Information Systems
Ashish Nainwal, Yatindra Kumar, Bhola Jha
Summary: Heart attack is a leading cause of death globally, and heart-related diseases have resulted in increased healthcare expenditure. This study proposes a method for diagnosing arrhythmia based on electrocardiogram signals, utilizing long-term recording devices to capture rare abnormal events. By optimizing feature vectors and applying convolutional neural networks for signal classification, the proposed method achieves comparably high classification accuracy.
JOURNAL OF KING SAUD UNIVERSITY-COMPUTER AND INFORMATION SCIENCES
(2022)
Article
Chemistry, Multidisciplinary
Junjie Zhou, Siyue Shuai, Lingyun Wang, Kaifeng Yu, Xiangjie Kong, Zuhua Xu, Zhijiang Shao
Summary: With the continuous development of smart cities, intelligent transportation systems have made many breakthroughs and upgrades. Traffic flow prediction plays a crucial role in managing complex traffic flow. However, existing methods often overlook the differences between lanes, resulting in less specific predictions. In order to address this issue, we propose a heterogeneous graph convolution model based on dynamic graph generation, which improves prediction accuracy by considering real-time spatial dependencies between lanes and other parameters.
APPLIED SCIENCES-BASEL
(2022)
Article
Plant Sciences
Jitong Cai, Renyong Pan, Jianwu Lin, Jiaming Liu, Licai Zhang, Xingtian Wen, Xiaoyulong Chen, Xin Zhang
Summary: This study proposes a model called FCA-EfficientNet, which demonstrates excellent performance in corn disease recognition. By introducing a fully-convolution-based coordinate attention module and an adaptive fusion module, the model can accurately identify diseases in complex backgrounds, improving the accuracy and real-time capability of corn disease recognition.
FRONTIERS IN PLANT SCIENCE
(2023)
Article
Biology
Matthew P. Adams, Arman Rahmim, Jing Tang
Summary: Integrating DAT SPECT images with UPDRS_III scores for deep learning prediction significantly improves outcome accuracy, allowing for easier and more universal application without the need for segmentation and feature extraction.
COMPUTERS IN BIOLOGY AND MEDICINE
(2021)
Article
Chemistry, Multidisciplinary
Hui Gao, Miaolei Deng, Wenjun Zhao, Dexian Zhang
Summary: A new crowd counting method called SASNet is proposed, which focuses on estimating crowd density in population heterogeneous distribution. The method utilizes scene adaptive segmentation network and dual branches network to achieve stabilized performance and robustness.
APPLIED SCIENCES-BASEL
(2022)
Article
Computer Science, Artificial Intelligence
Xiaoyu Kong, Ke Zhang
Summary: Human behavior is influenced by emotions, and predicting behavior through emotion classification from text is significant for decision-making. Efficiently extracting emotional tendencies from text data is a challenge, but a upgraded CNN model proposed in this study improves the downsides and shows better performance in sentiment analysis tasks.
PEERJ COMPUTER SCIENCE
(2023)