Article
Biochemistry & Molecular Biology
Victor Akpokiro, H. M. A. Mohit Chowdhury, Samuel Olowofila, Raisa Nusrat, Oluwatosin Oluwadare
Summary: The article introduces CNNSplice, a set of deep convolutional neural network models for splice site prediction. Through model selection technique, the authors propose five high-performing models that efficiently predict true and false splice sites in balanced and imbalanced datasets. Evaluation results indicate that CNNSplice's models achieve better performance compared with existing methods across five organisms' datasets and also demonstrate generalizability on new or poorly trained genome datasets.
COMPUTATIONAL AND STRUCTURAL BIOTECHNOLOGY JOURNAL
(2023)
Article
Biochemical Research Methods
Hasan Zulfiqar, Zi-Jie Sun, Qin-Lai Huang, Shi-Shi Yuan, Hao Lv, Fu-Ying Dao, Hao Lin, Yan-Wen Li
Summary: This study developed a deep learning-based model to predict 4mC sites in Escherichia coli. By encoding DNA sequences and utilizing convolutional neural networks for classification, the model can accurately identify 4mC sites, providing a convenient approach for studying 4mC modification.
Article
Biochemical Research Methods
Victor Akpokiro, Trevor Martin, Oluwatosin Oluwadare
Summary: EnsembleSplice, an ensemble learning architecture consisting of four distinct convolutional neural networks (CNN), outperforms existing splice site detection methods. Through five-fold cross-validation, the best-performing model was identified based on evaluation and diversity metrics, leading to the development of an effective splice site detection model.
BMC BIOINFORMATICS
(2022)
Article
Computer Science, Artificial Intelligence
Erfan A. Shams, Ahmet Rizaner, Ali Hakan Ulusoy
Summary: A new context-aware feature extraction method was proposed for CNN-based multiclass intrusion detection, which effectively improved classification accuracy by reducing feature space and classification time. The study showed that the method performed well on multiple datasets and enhanced the performance of intrusion detection.
NEURAL COMPUTING & APPLICATIONS
(2021)
Article
Biology
Md Rabiul Islam, Md Milon Islam, Md Mustafizur Rahman, Chayan Mondal, Suvojit Kumar Singha, Mohiuddin Ahmad, Abdul Awal, Md Saiful Islam, Mohammad Ali Moni
Summary: The study proposed a deep machine-learning model using Convolutional Neural Network to convert EEG data and recognize emotions on images, overcoming the challenge of emotion recognition from low amplitude variation in EEG signals.
COMPUTERS IN BIOLOGY AND MEDICINE
(2021)
Article
Multidisciplinary Sciences
Yuanyuan Qu, Xuesheng Li, Zhiliang Qin, Qidong Lu
Summary: In this paper, a novel approach based on a multi-branch three-dimensional (3D) convolution neural network (CNN) model is proposed for accurate acoustic scene classification (ASC). Multiple frequency-domain representations of signals are formed by utilizing expert knowledge on acoustics and discrete wavelet transformations (DWT). The proposed 3D CNN architecture, featuring residual connections and squeeze-and-excitation attentions (3D-SE-ResNet), effectively captures both long-term and short-term correlations in environmental sounds. Additionally, an auxiliary supervised branch based on the chromatogram of the original signal is incorporated to alleviate overfitting risks. Numerical evaluation on a large-scale dataset demonstrates the superior performance of the proposed multi-input multi-feature 3D-CNN architecture over state-of-the-art methods.
SCIENTIFIC REPORTS
(2022)
Article
Engineering, Biomedical
Rohollah Hedayati, Mohammad Khedmati, Mehran Taghipour-Gorjikolaie
Summary: Alzheimer's disease is a major cause of death among the elderly, but early diagnosis is difficult. Machine learning methods have been used to improve accuracy in diagnosing Alzheimer's disease, with results showing accuracy rates of 95% for AD/NC, 90% for AD/MCI, and 92.5% for MCI/NC. This indicates the method is reliable for early diagnosis of Alzheimer's disease.
BIOMEDICAL SIGNAL PROCESSING AND CONTROL
(2021)
Article
Computer Science, Artificial Intelligence
Jing Yuan, Shuwei Cao, Gangxing Ren, Fengxian Su, Huiming Jiang, Qian Zhao
Summary: In this paper, a new mechanical feature extraction and fault diagnosis method LW-Net is introduced, which achieves accurate extraction of impact fault features and improves fault diagnosis effectiveness through smart lifting wavelet kernels and the design of lifting layer.
NEURAL COMPUTING & APPLICATIONS
(2022)
Article
Remote Sensing
Wangbin Li, Kaimin Sun, Hepeng Zhao, Wenzhuo Li, Jinjiang Wei, Song Gao
Summary: Building extraction from remote sensing imagery is a common task in surveying, mapping and geographic information systems. This study proposes a pyramid feature extraction method to address the challenges of automatic building extraction. The method constructs multi-scale representations of buildings and incorporates attention modules and feature alignment modules to improve the accuracy and integrity of the extraction results.
INTERNATIONAL JOURNAL OF APPLIED EARTH OBSERVATION AND GEOINFORMATION
(2022)
Article
Biochemical Research Methods
Yongxian Fan, Guicong Sun, Xiaoyong Pan
Summary: In this study, a contextual language embedding-based method called ELMo4m6A was proposed to predict m6A sites from RNA sequences without any prior knowledge. The method first learns embeddings of RNA sequences using the language model ELMo, and then utilizes a hybrid CNN and LSTM to identify m6A sites. Results from 5-fold cross-validation and independent testing showed that ELMo4m6A outperforms state-of-the-art methods. Additionally, integrated gradients were applied to identify potential sequence patterns contributing to m6A sites.
IEEE-ACM TRANSACTIONS ON COMPUTATIONAL BIOLOGY AND BIOINFORMATICS
(2023)
Article
Engineering, Biomedical
M. A. H. Akhand, Mahfuza Akter Maria, Md Abdus Samad Kamal, Tetsuya Shimamura
Summary: The study proposes an enhanced connectivity feature map for emotion recognition by introducing partial mutual information and an additional channel, which improves the performance of emotion recognition by extracting more information from brain signals.
BIOMEDICAL SIGNAL PROCESSING AND CONTROL
(2024)
Article
Multidisciplinary Sciences
Xiaolei Wang, Zirong Hu, Shouhai Shi, Mei Hou, Lei Xu, Xiang Zhang
Summary: In this paper, a convolutional network called Adaptive Feature Fusion UNet (AFF-UNet) was proposed to optimize the semantic segmentation performance of remote sensing imagery (RSI). The AFF-UNet model consists of dense skip connections architecture, an adaptive feature fusion module, a channel attention convolution block, and a spatial attention module. Experimental results showed that the proposed model achieved improvements in average F1 score and overall accuracy compared to DeepLabv3+, demonstrating better segmentation performance and object integrity.
SCIENTIFIC REPORTS
(2023)
Article
Thermodynamics
Maryam Imani
Summary: A new nonlinear relationship extraction method is proposed in this work, using convolutional neural network and support vector regression for load forecasting, showing superior performance compared to several outstanding forecasters.
Article
Physics, Multidisciplinary
Junjiang Zhu, Jintao Lv, Dongdong Kong
Summary: The study proposes an improved deep learning method, CNN-FWS, for accurate diagnosis of abnormal and normal ECG signals. The method combines three convolutional neural networks and recursive feature elimination based on feature weights. The experimental results demonstrate high F1 score and Recall values for the CNN-FWS model, showing its effectiveness in diagnosing abnormal ECG signals.
Article
Biochemical Research Methods
Fang Jing, Shao-Wu Zhang, Shihua Zhang
Summary: In this study, a new method (ACNN) was proposed to predict transcription factor binding sites (TFBSs), effectively capturing global and local contexts in DNA sequences and features of histone modification markers. Through adversarial training, ACNN is able to learn common features and improve prediction performance, and performs better than a baseline method with limited labeled data.
IEEE-ACM TRANSACTIONS ON COMPUTATIONAL BIOLOGY AND BIOINFORMATICS
(2022)
Article
Computer Science, Information Systems
Chenchu Xu, Yifei Wang, Dong Zhang, Longfei Han, Yanping Zhang, Jie Chen, Shuo Li
Summary: This paper proposes a semi-supervised myocardial infarction segmentation method that consists of a boundary mining model and an adversarial learning model. The boundary mining model solves the boundary ambiguity problem by enlarging the gap between foreground and background features. The adversarial learning model enables the boundary mining model to learn from additional unlabeled data, increasing the model's robustness.
IEEE JOURNAL OF BIOMEDICAL AND HEALTH INFORMATICS
(2023)
Article
Computer Science, Information Systems
Xingxin Xu, Qikui Zhu, Hanning Ying, Jiongcheng Li, Xiujun Cai, Shuo Li, Xiaoqing Liu, Yizhou Yu
Summary: In this study, a Knowledge-guided framework named MCCNet is proposed to adaptively integrate multi-phase liver lesion information and construct a lesion classification network. The effectiveness of the proposed modules in exploiting and fusing multi-phase information is demonstrated through extensive experimental results and evaluations on a dataset containing 3,683 lesions from 2,333 patients in 9 hospitals.
IEEE JOURNAL OF BIOMEDICAL AND HEALTH INFORMATICS
(2023)
Review
Computer Science, Artificial Intelligence
Chenji Zhao, Shun Xiang, Yuanquan Wang, Zhaoxi Cai, Jun Shen, Shoujun Zhou, Di Zhao, Weihua Su, Shijie Guo, Shuo Li
Summary: This paper proposes a context-aware network (CA-Net) for semi-supervised left atrium (LA) segmentation from 3D MRI. Experimental results show that contextual information is helpful in extracting accurate atrial structures, and CA-Net achieves better performance than some state-of-the-art semi-supervised networks.
EXPERT SYSTEMS WITH APPLICATIONS
(2023)
Article
Computer Science, Artificial Intelligence
Ziyue Jiang, Yuting He, Shuai Ye, Pengfei Shao, Xiaomei Zhu, Yi Xu, Yang Chen, Jean-Louis Coatrieux, Shuo Li, Guanyu Yang
Summary: This paper proposes a one-to-multiple unsupervised domain adaptation method for medical image segmentation. The proposed method utilizes dynamic domain adaptation and hybrid uncertainty learning to improve the model's generalization ability. Experimental results demonstrate the effectiveness of the proposed method in achieving competitive segmentation results and high adaptation.
KNOWLEDGE-BASED SYSTEMS
(2023)
Article
Computer Science, Artificial Intelligence
Dong Zhang, Chenchu Xu, Shuo Li
Summary: This study proposes the Heuristic Multi-modal Integration (HMI) framework to detect liver tumors from multi-modal non-enhanced MRIs. The HMI utilizes individual DRL modules on each modality to extract specific features and then integrates these modules into a collective DRL module, utilizing the comprehensive information from multiple modalities to detect the desired tumors. The experimental results show that the HMI outperforms current state-of-the-art methods, making it an accurate and contrast-agent-free alternative for liver tumor detection in clinical settings.
EXPERT SYSTEMS WITH APPLICATIONS
(2023)
Article
Computer Science, Artificial Intelligence
Shen Zhao, Jinhong Wang, Xinxin Wang, Yikang Wang, Hanying Zheng, Bin Chen, An Zeng, Fuxin Wei, Sadeer Al-Kindi, Shuo Li
Summary: Automatic vertebral body contour extraction (AVBCE) is essential for the diagnosis and treatment of spinal diseases. This study proposes a novel method called ADMIRE, which extends the target contour capture range and provides morphology-aware parameters. Experimental results demonstrate that ADMIRE achieves state-of-the-art performance on various datasets, indicating its accuracy, robustness, and generalization ability.
MEDICAL IMAGE ANALYSIS
(2023)
Article
Computer Science, Information Systems
Gongning Luo, Xinghua Ma, Jinwen Guo, Mingye Zou, Wei Wang, Yang Cao, Kuanquan Wang, Shuo Li
Summary: This paper introduces a method for removing guidewire artifacts in IVOCT videos and proposes a Trajectory-aware Adaptive imaging Clue analysis Network (TAC-Net). TAC-Net, with its innovative designs of adaptive clue aggregation and trajectory-aware transformer, reliably restores the texture and structure of artifact areas.
IEEE JOURNAL OF BIOMEDICAL AND HEALTH INFORMATICS
(2023)
Article
Computer Science, Information Systems
Kaini Wang, Shuaishuai Zhuang, Juzheng Miao, Yang Chen, Jie Hua, Guang-Quan Zhou, Xiaopu He, Shuo Li
Summary: This article proposes a framework for comprehensive learning in the frequency domain to identify colonic disease subtypes. The experimental results demonstrate that the proposed method achieves state-of-the-art accuracy rate.
IEEE JOURNAL OF BIOMEDICAL AND HEALTH INFORMATICS
(2023)
Article
Computer Science, Information Systems
Yinli Tian, Wenjian Qin, Fei Xue, Ricardo Lambo, Meiyan Yue, Songhui Diao, Lequan Yu, Yaoqin Xie, Hailin Cao, Shuo Li
Summary: In this paper, a novel fine-grained segmentation framework called ARR-GCN is proposed, which incorporates prior anatomical relations into the learning process. The framework outperforms other methods in liver segment and lung lobe segmentation tasks.
IEEE JOURNAL OF BIOMEDICAL AND HEALTH INFORMATICS
(2023)
Article
Automation & Control Systems
Liansheng Wang, Jiacheng Wang, Lei Zhu, Huazhu Fu, Ping Li, Gary Cheng, Zhipeng Feng, Shuo Li, Pheng-Ann Heng
Summary: Automated detection of lung infections from CT data is crucial for combatting COVID-19. However, there are challenges in AI system development, such as the reliance on 2D CT images, limitations of existing 3D segmentation methods, and the lack of annotated CT volumes. To address these issues, a multiple dimensional-attention CNN is proposed to aggregate multiscale information, and a dual multiscale mean teacher network is utilized for semi-supervised segmentation. Experimental results demonstrate the superiority of this approach over state-of-the-art methods.
IEEE TRANSACTIONS ON CYBERNETICS
(2023)
Article
Engineering, Electrical & Electronic
Shimeng Yang, Teng Li, Yinping Lv, Yi Xia, Shuo Li
Summary: This letter proposes a boundary-guided pseudo-labeling method that utilizes prior anatomical knowledge to generate and select reliable pseudo-labels for unlabeled data to improve measurement performance. The proposed method incorporates a fine self-attention module and a boundary attention module to enhance the quality of pseudo-labels. Experiments conducted on challenging carotid ultrasound datasets demonstrate that the proposed method outperforms existing state-of-the-art algorithms.
IEEE SIGNAL PROCESSING LETTERS
(2023)
Article
Computer Science, Artificial Intelligence
Zhongyi Han, Xian-Jin Gui, Haoliang Sun, Yilong Yin, Shuo Li
Summary: In this paper, a noise-robust domain adaptation method is proposed to address the issue of corrupted source domain examples in multiple noisy environments. By utilizing offline curriculum learning, gradually decreasing noisy distribution distance, estimating open-set noise degree, robust parameter learning, and domain-invariant feature learning, these components are seamlessly transformed into an adversarial network for efficient joint optimization, leading to significant improvements in transfer tasks.
IEEE TRANSACTIONS ON PATTERN ANALYSIS AND MACHINE INTELLIGENCE
(2023)
Article
Computer Science, Information Systems
Guanyu Yang, Yuting He, Yang Lv, Yang Chen, Jean-Louis Coatrieux, Xiaoxuan Sun, Qiang Wang, Yongyue Wei, Shuo Li, Yinsu Zhu
Summary: PAH treatment requires accurate prognosis prediction on 3D non-contrast CT images, which is challenging due to the large volume and low contrast regions. In this paper, we propose a multi-task learning-based framework, P-2-Net, which optimizes the model and represents task-dependent features effectively. Our approach utilizes Memory Drift (MD) to densely sample biomarkers and Prior Prompt Learning (PPL) to embed clinical prior knowledge, achieving high prognostic correlation and great generalization.
IEEE JOURNAL OF BIOMEDICAL AND HEALTH INFORMATICS
(2023)
Article
Computer Science, Artificial Intelligence
Chenchu Xu, Yuhong Song, Dong Zhang, Leonardo Kayat Bittencourt, Sree Harsha Tirumani, Shuo Li
Summary: Liver tumor detection without contrast agents has great potential in advancing liver cancer screening. This paper proposes a novel teacher-student reinforcement learning method that allows the student network to directly detect tumors from non-enhanced images guided by explicit liver tumor knowledge obtained from contrast-enhanced images. Experimental results show that this method achieves state-of-the-art performance in liver tumor detection without contrast agents.
MEDICAL IMAGE ANALYSIS
(2023)
Article
Computer Science, Artificial Intelligence
Xiangyu Li, Xinjie Liang, Gongning Luo, Wei Wang, Kuanquan Wang, Shuo Li
Summary: In this study, we propose a novel framework based on label distribution learning (LDL) paradigm to estimate tumor cellularity (TC). The proposed framework addresses the challenges of inter-rater ambiguity exploitation, label distribution generation, and accurate TC value recovery. It achieves superior performance compared to existing methods on the TC estimation task.
MEDICAL IMAGE ANALYSIS
(2023)
