Article
Computer Science, Artificial Intelligence
Lei Bi, Michael Fulham, Jinman Kim
Summary: Segmentation of skin lesions using fully convolutional networks is accurate but limited by insufficient training data. Semi-automatic segmentation methods that combine user-inputs with high-level image features offer a better solution for challenging skin lesions. The proposed hyper-fusion network (HFN) in this study outperformed state-of-the-art methods in accuracy and generalizability.
MEDICAL IMAGE ANALYSIS
(2022)
Article
Computer Science, Artificial Intelligence
Zhengeng Yang, Hongshan Yu, Yong He, Wei Sun, Zhi-Hong Mao, Ajmal Mian
Summary: This article focuses on learning representation from unlabeled data for semantic segmentation. A novel self-supervised learning framework is developed by formulating the jigsaw puzzle problem as a patch-wise classification problem and solving it with a fully convolutional network. Significant improvements are achieved on the Cityscapes dataset and competitive performance is demonstrated on the PASCAL VOC2012 dataset with fewer time costs on pretraining.
IEEE TRANSACTIONS ON NEURAL NETWORKS AND LEARNING SYSTEMS
(2022)
Article
Environmental Sciences
Guanzhou Chen, Xiaoliang Tan, Beibei Guo, Kun Zhu, Puyun Liao, Tong Wang, Qing Wang, Xiaodong Zhang
Summary: Semantic segmentation is a fundamental task in remote sensing image analysis, and our proposed SDFCNv2 framework shows better performance on remote sensing images compared to the SDFCNv1 framework, increasing the mIoU metric by up to 5.22% while using only about half of the parameters.
Article
Computer Science, Artificial Intelligence
Jun Zhang, Zhiyuan Hua, Kezhou Yan, Kuan Tian, Jianhua Yao, Eryun Liu, Mingxia Liu, Xiao Han
Summary: This paper introduces a weakly-supervised model using joint fully convolutional and graph convolutional networks for automated segmentation of pathology images. By utilizing image-level labels instead of pixel-wise annotations, the segmentation model's performance is improved. Experimental results demonstrate the effectiveness of this method in cancer region segmentation.
MEDICAL IMAGE ANALYSIS
(2021)
Article
Computer Science, Interdisciplinary Applications
Lei Bi, Michael Fulham, Nan Li, Qiufang Liu, Shaoli Song, David Dagan Feng, Jinman Kim
Summary: A recurrent fusion network (RFN) is proposed in this study to progressively fuse multi-modality image features through multiple recurrent fusion phases, producing consistent segmentation results across different network architectures. The RFN method shows more accurate segmentation compared to existing methods and is generalizable to different datasets.
COMPUTER METHODS AND PROGRAMS IN BIOMEDICINE
(2021)
Article
Multidisciplinary Sciences
Nayansi Jha, Taehun Kim, Sungwon Ham, Seung-Hak Baek, Sang-Jin Sung, Yoon-Ji Kim, Namkug Kim
Summary: The aim of this study was to develop an auto-segmentation algorithm for mandibular condyle using the 3D U-Net and determine the optimal dataset size for achieving clinically acceptable accuracy through a stress test. The results showed that increasing the training data improves the segmentation accuracy for mandibular condyle fractures.
SCIENTIFIC REPORTS
(2022)
Article
Engineering, Biomedical
Dali F. D. dos Santos, Paulo R. de Faria, Bruno A. N. Travencolo, Marcelo Z. do Nascimento
Summary: The study proposes a method based on a fully convolutional neural network for localizing and performing refined segmentation of tumor regions in histological whole slide images. Experimental results show that the method achieved good results in different cancer-derived datasets with high accuracy up to 97.6%.
BIOMEDICAL SIGNAL PROCESSING AND CONTROL
(2021)
Article
Geochemistry & Geophysics
Guanzhou Chen, Chanjuan He, Tong Wang, Kun Zhu, Puyun Liao, Xiaodong Zhang
Summary: This article introduces an efficient unsupervised remote sensing image segmentation method based on superpixel segmentation and fully convolutional networks. The method can rapidly achieve pixel-level image segmentation without requiring manual labels or prior knowledge.
IEEE GEOSCIENCE AND REMOTE SENSING LETTERS
(2022)
Article
Engineering, Electrical & Electronic
Hai Huang, Chang Liu, Lei Tian, Junsheng Mu, Xiaojun Jing
Summary: The proposed video saliency detection model enhances object contour depiction by introducing a ConvLSTM module, allowing it to learn spatial and temporal information. Through the use of augmentation techniques to expand the dataset, and training and evaluation on widely used datasets, the model demonstrates improved performance in detecting moving objects.
INTERNATIONAL JOURNAL OF CIRCUIT THEORY AND APPLICATIONS
(2021)
Article
Acoustics
Xin Jia, Xiejing Li, Ting Shen, Ling Zhou, Guang Yang, Fan Wang, Xingguang Zhu, Mingxi Wan, Shiyan Li, Siyuan Zhang
Summary: This study innovatively applied fully convolutional neural networks (FCNs) for detection and monitoring of thermal ablation regions in ultrasound (US) and comprehensively compared the performance of several models in medical image segmentation. The UNet++ and ResUNet showed relatively outstanding segmentation performance among all compared models, making them potential candidates for automatic segmentation of thermal ablation regions in US during clinical ablation treatment.
Article
Chemistry, Analytical
Chih-Chiang Wei, Tzu-Heng Huang
Summary: The study utilized fully convolutional networks to establish a forecasting model for predicting hourly rainfall during typhoons in Taiwan. By deep learning and image recognition technology, the model effectively improved the accuracy of rainfall forecasting during typhoons in southern Taiwan.
Article
Engineering, Multidisciplinary
Qian Wu, Jinan Gu, Chen Wu, Jin Li
Summary: This paper proposes a CRF-FCN model based on CRF optimization to improve semantic segmentation results by making target details more obvious, and enhance the accuracy and efficiency of semantic segmentation.
JOURNAL OF COMPUTATIONAL METHODS IN SCIENCES AND ENGINEERING
(2021)
Article
Computer Science, Information Systems
Jian Ji, Xiaocong Lu, Mai Luo, Minghui Yin, Qiguang Miao, Xiangzeng Liu
Summary: This study proposed a parallel fully convolutional neural network that integrated edge detection to improve semantic segmentation performance. Through comprehensive experiments on multiple datasets, better results were achieved and compared with other methods.
Article
Computer Science, Information Systems
Hiba Mzoughi, Ines Njeh, Mohamed Ben Slima, Ahmed Ben Hamida, Chokri Mhiri, Kheireddine Ben Mahfoudh
Summary: This paper investigates a fully automatic Computer-Aided Diagnosis (CAD) tool for brain glioblastomas tumor exploration based on convolutional Deep-Learning algorithms. The CAD tool includes three steps: pre-processing, segmentation, and classification. Experimental results demonstrate the efficiency and accuracy of the CAD tool in diagnosing brain glioblastomas tumors, highlighting its significance in improving diagnostic performance.
MULTIMEDIA TOOLS AND APPLICATIONS
(2021)
Article
Computer Science, Artificial Intelligence
Quan Zhou, Xiaofu Wu, Suofei Zhang, Bin Kang, Zongyuan Ge, Longin Jan Latecki
Summary: This paper introduces a novel encoder-decoder architecture called CENet for semantic segmentation, which achieves superior performance on two widely-used semantic segmentation datasets and obtains promising results on instance segmentation and biological segmentation tasks.
PATTERN RECOGNITION
(2022)
Article
Clinical Neurology
Patrick H. Luckett, Charlie Chen, Brian A. Gordon, Julie Wisch, Sarah B. Berman, Jasmeer P. Chhatwal, Carlos Cruchaga, Anne M. Fagan, Martin R. Farlow, Nick C. Fox, Mathias Jucker, Johannes Levin, Colin L. Masters, Hiroshi Mori, James M. Noble, Stephen Salloway, Peter R. Schofield, Adam M. Brickman, William S. Brooks, David M. Cash, Michael J. Fulham, Bernardino Ghetti, Clifford R. Jack, Jonathan Voeglein, William E. Klunk, Robert Koeppe, Yi Su, Michael Weiner, Qing Wang, Daniel Marcus, Deborah Koudelis, Nelly Joseph-Mathurin, Lisa Cash, Russ Hornbeck, Chengjie Xiong, Richard J. Perrin, Celeste M. Karch, Jason Hassenstab, Eric McDade, John C. Morris, Tammie L. S. Benzinger, Randall J. Bateman, Beau M. Ances
Summary: This study analyzed 19 biomarkers of Alzheimer's disease using hierarchical clustering and feature selection, and found that amyloid and tau measures were the primary predictors. Emerging biomarkers of neuronal integrity and inflammation showed weaker predictive ability.
ALZHEIMERS & DEMENTIA
(2023)
Editorial Material
Clinical Neurology
Sophie Dunkerton, Ross Penninkilampi, Heidi Beadnall, Michael Fulham, Andrew Colebatch, Stacey Jankelowitz, Rebekah Ahmed, Zoe Thayer, Michael Halmagyi, Edward Abadir
PRACTICAL NEUROLOGY
(2023)
Article
Computer Science, Interdisciplinary Applications
Xiaohang Fu, Lei Bi, Ashnil Kumar, Michael Fulham, Jinman Kim
Summary: The identification of melanoma can be done through the analysis of clinical and dermoscopy images. Current methods lack the ability to fully utilize information from both modalities and exploit the intercategory relationships in the 7PC. This study proposes a graph-based network with two modules to address these limitations and improves classification performance.
IEEE TRANSACTIONS ON MEDICAL IMAGING
(2022)
Article
Biology
Xiaohang Fu, Ellis Patrick, Jean Y. H. Yang, David Dagan Feng, Jinman Kim
Summary: The spatial architecture and phenotypic heterogeneity of tumor cells are associated with cancer prognosis and outcomes. Imaging mass cytometry captures high-dimensional maps of disease-relevant biomarkers at single-cell resolution, which can inform patient-specific prognosis. However, existing methods for survival prediction do not utilize spatial phenotype information at the single-cell level, and there is a lack of end-to-end methods that integrate imaging data with clinical information for improved accuracy. We propose a deep multimodal graph-based network that considers spatial phenotype information and clinical variables to enhance survival prediction, and demonstrate its effectiveness in breast cancer datasets.
COMPUTERS IN BIOLOGY AND MEDICINE
(2023)
Article
Computer Science, Information Systems
Usman Naseem, Matloob Khushi, Jinman Kim
Summary: Pathology visual question answering (PathVQA) aims to answer medical questions using pathology images. Existing methods have limitations in capturing the high and low-level interactions between vision and language features required for VQA. Additionally, these methods lack interpretability in justifying the retrieved answers. To address these limitations, a vision-language transformer called TraP-VQA is introduced, which embeds vision and language features for interpretable PathVQA. Our experiments demonstrate that TraP-VQA outperforms state-of-the-art methods and validate its robustness on medical VQA datasets, along with the capability of the integrated vision-language model. Visualization results explain the reasoning behind the retrieved PathVQA answers.
IEEE JOURNAL OF BIOMEDICAL AND HEALTH INFORMATICS
(2023)
Article
Automation & Control Systems
Yuyu Guo, Lei Bi, Dongming Wei, Liyun Chen, Zhengbin Zhu, Dagan Feng, Ruiyan Zhang, Qian Wang, Jinman Kim
Summary: In this study, we propose a dense-sparse-dense (DSD) motion estimation framework that utilizes unsupervised 3D landmark detection network and motion reconstruction network to extract sparse landmarks and construct motion field in two stages. The method improves the accuracy of motion estimation and preserves anatomical topology.
IEEE TRANSACTIONS ON CYBERNETICS
(2023)
Letter
Hematology
Judith Trotman, Peter Presgrave, Duncan P. Carradice, Douglas Stuart Lenton, Maher K. Gandhi, Tara Cochrane, Xavier Badoux, Julia Carlson, Gloria Nkhoma, Belinda Butcher, Armin Nikpour, Michael Fulham, Anna M. Johnston
Article
Computer Science, Interdisciplinary Applications
Wenxiang Ding, Qiaoqiao Ding, Kewei Chen, Miao Zhang, Li Lv, David Dagan Feng, Lei Bi, Jinman Kim, Qiu Huang
Summary: Dynamic PET imaging provides more comprehensive physiological information than conventional static PET imaging. The proposed modified Logan reference plot model and self-supervised convolutional neural network improve noise performance and accurately estimate the distribution volume ratio in dynamic PET with a shortened scanning protocol. The method has the potential to add clinical value by providing both DVR and SUV simultaneously.
IEEE TRANSACTIONS ON MEDICAL IMAGING
(2023)
Article
Computer Science, Cybernetics
Usman Naseem, Matloob Khushi, Jinman Kim, Adam G. Dunn
Summary: People on social media using disease and symptom words to discuss their health can introduce biases in data-driven public health applications. This study presents a new dataset called RHMD, which consists of 10,015 manually annotated Reddit posts. The dataset is labeled with four categories and provides a comprehensive performance analysis of baseline methods. The release of this dataset is expected to facilitate the development of new methods for detecting health mentions in user-generated text.
IEEE TRANSACTIONS ON COMPUTATIONAL SOCIAL SYSTEMS
(2023)
Review
Automation & Control Systems
Wei-Chien Wang, Euijoon Ahn, Dagan Feng, Jinman Kim
Summary: Over the last decade, supervised deep learning has made significant progress in computer vision tasks using manually annotated big data. However, the limited availability of high-quality annotated medical imaging data hinders the application of deep learning in medical image analysis. A potential solution is the use of self-supervised learning (SSL), particularly contrastive SSL, which has shown promise in rivaling or surpassing supervised learning. This review examines state-of-the-art contrastive SSL algorithms originally designed for natural images, explores their adaptations for medical images, and discusses recent advances, current limitations, and future directions in applying contrastive SSL in the medical domain.
MACHINE INTELLIGENCE RESEARCH
(2023)
Article
Biochemistry & Molecular Biology
Hendris Wongso, Maiko Ono, Tomoteru Yamasaki, Katsushi Kumata, Makoto Higuchi, Ming-Rong Zhang, Michael J. Fulham, Andrew Katsifis, Paul A. Keller
Summary: The pyridinyl-butadienyl-benzothiazole (PBB3 15) scaffold was used to improve tau ligands for imaging Alzheimer's disease. Triazole derivatives visualized A beta plaques but failed to detect neurofibrillary tangles (NFTs), while amide 110 and ester 129 successfully observed NFTs. These ligands showed different affinities at the binding sites with PBB3.
RSC MEDICINAL CHEMISTRY
(2023)
Article
Computer Science, Cybernetics
Usman Naseem, Matloob Khushi, Jinman Kim, Adam G. Dunn
Summary: The article introduces a hybrid text representation method for explaining suicide risk identification on social media. The method achieves excellent results on a public suicide dataset and demonstrates advantages in clinical and public health practice.
IEEE TRANSACTIONS ON COMPUTATIONAL SOCIAL SYSTEMS
(2022)
Article
Computer Science, Artificial Intelligence
Ahmad Karambakhsh, Bin Sheng, Ping Li, Huating Li, Jinman Kim, Younhyun Jung, C. L. Philip Chen
Summary: The article introduces a novel solution for 3-D object recognition from volumetric data by combining three compact CNN models, low-cost SparseNet, and feature representation technique. By estimating extra geometrical information, an optimized network is achieved and improves the recognition results.
IEEE TRANSACTIONS ON NEURAL NETWORKS AND LEARNING SYSTEMS
(2022)