Article
Biology
Sanguk Park, Minyoung Chung
Summary: The segmentation of cardiac structures in CT images is crucial for diagnosing cardiovascular diseases. This study introduces a novel model focusing on shape and boundary-aware features to improve accuracy between proximate organs. The proposed network outperforms state-of-the-art models by enhancing the attention on edges between substructures.
COMPUTERS IN BIOLOGY AND MEDICINE
(2022)
Article
Multidisciplinary Sciences
Yongning Zou, Gongjie Yao, Jue Wang
Summary: In this paper, a framework for CT image segmentation of oil rock core is proposed. The improved level set segmentation algorithm, based on the characteristics of CT images of oil rock core, has been applied with promising results.
Article
Computer Science, Artificial Intelligence
Xiaoming Liu, Quan Yuan, Yaozong Gao, Kelei He, Shuo Wang, Xiao Tang, Jinshan Tang, Dinggang Shen
Summary: This paper proposed a novel weakly supervised segmentation method for COVID-19 infections in CT slices, which only requires scribble supervision and is enhanced with the uncertainty-aware self-ensembling and transformation consistent techniques. Experimental results demonstrate that the proposed method is more effective than other weakly supervised methods and achieves similar performance as those fully supervised.
PATTERN RECOGNITION
(2022)
Article
Computer Science, Artificial Intelligence
Xiaoyu Chen, Hong-Yu Zhou, Feng Liu, Jiansen Guo, Lianshen Wang, Yizhou Yu
Summary: This paper proposes a modality-collaborative semi-supervised segmentation (MASS) method for medical images, which utilizes modality-independent knowledge learned from unpaired CT and MRI scans. By employing cross-modal consistency and contrastive similarity loss, MASS regularizes deep segmentation models to achieve improved segmentation results under extremely limited supervision.
MEDICAL IMAGE ANALYSIS
(2022)
Article
Computer Science, Artificial Intelligence
Aldimir Bruzadin, Maurilio Boaventura, Marilaine Colnago, Rogerio Galante Negri, Wallace Casaca
Summary: Deep Learning is a key approach for medical imaging challenges, including lung segmentation in CT. This paper proposes a semi-automatic framework that combines deep contour learning with seeded segmentation to accurately segment lung CT images from COVID-19 patients, even in the absence of well-defined boundaries and regardless of infection level.
Article
Biology
Zhengshan Huang, Yu Guo, Ning Zhang, Xian Huang, Pierre Decazes, Stephanie Becker, Su Ruan
Summary: Accurate lymphoma segmentation in PET/CT images is crucial for DLBCL prognosis evaluation. This study proposes a weakly supervised deep learning method based on multi-scale feature similarity for automatic lymphoma segmentation, which effectively reduces the reliance on accurately labeled datasets and expert annotations.
COMPUTERS IN BIOLOGY AND MEDICINE
(2022)
Article
Radiology, Nuclear Medicine & Medical Imaging
Chengyan Yuan, Shuni Song, Jinzhong Yang, Yu Sun, Benqiang Yang, Lisheng Xu
Summary: This study aims to improve the accuracy of pulmonary artery segmentation by proposing a new network model and loss function, and using deep learning methods. The experimental results demonstrate a significant improvement in segmentation accuracy, making it suitable for the rapid diagnosis of pulmonary artery diseases.
Article
Computer Science, Artificial Intelligence
Caixia Liu, Wanli Xie, Ruibin Zhao, Mingyong Pang
Summary: This article presents an automatic algorithm for accurately segmenting lungs from thoracic CT images, which includes three principal steps: image preprocessing, lung extracting, and contour correcting. Experimental results show that the algorithm outperforms other methods on image denoising and smoothing, and achieves a high segmentation accuracy on a group of lung CT images affected with interstitial lung diseases.
IET IMAGE PROCESSING
(2023)
Article
Engineering, Civil
Nan Xiao, Li-Cheng Luo, Fu Huang, Tong-Hua Ling
Summary: This paper investigates the fracture characteristics of rock by pre-fabricating cracks with different angles in rock specimens and analyzing the effects on rock mechanical properties and crack propagation modes. The relationship between porosity and Young's modulus as well as fracture property is explored by numerical modeling using the cohesive element method and CT images. The results suggest that pre-fabricated cracks reduce the mechanical properties of rock and that the angles of the cracks affect the type of extended crack.
GEOMECHANICS AND ENGINEERING
(2022)
Article
Computer Science, Artificial Intelligence
Fengli Lu, Chengcai Fu, Guoying Zhang, Jie Shi
Summary: This paper proposes a novel adaptive multi-scale feature fusion method based on U-net (AMSFF-U-net) for fracture segmentation in coal rock CT images. The proposed method can effectively capture fracture features and achieve better segmentation performance, especially for weak fractures and small-scale fractures.
JOURNAL OF INTELLIGENT & FUZZY SYSTEMS
(2022)
Article
Computer Science, Information Systems
Fengli Lu, Chengcai Fu, Jie Shi, Guoying Zhang
Summary: This paper proposes a deep neural network (A-DNNet) based on the U-Net framework for extracting micro-fractures from sequential coal rock images. Using techniques such as CNN encoder, ConvLSTM, and RFA, this method effectively addresses the challenges of low contrast and high noise in fracture extraction. Experimental results demonstrate that the proposed method outperforms U-Net algorithms in terms of accuracy, precision, Dice coefficient, and F1 score.
MULTIMEDIA TOOLS AND APPLICATIONS
(2022)
Article
Energy & Fuels
Cun Zhang, Sheng Jia, Xuanhao Huang, Xutao Shi, Tong Zhang, Lei Zhang, Fangtian Wang
Summary: Pores, fractures, and mineral compositions in coal have a significant impact on the mechanical properties, seepage characteristics, and water-rock interaction of coal. This paper proposes a method to determine the threshold value of various mineral components and pores inside the coal sample through computed tomography (CT) scanning, X-ray diffractometer (XRD) test and nuclear magnetic resonance (NMR) technology.
Article
Chemistry, Multidisciplinary
Jianjun Peng, Yunhao Cui, Zhidan Zhong, Yi An
Summary: This paper proposes an ore rock fragmentation calculation method (ORFCM) based on the multi-modal fusion of point clouds and images. The ORFCM accurately calculates the ore rock fragmentation in the local excavation area and meets practical engineering needs.
APPLIED SCIENCES-BASEL
(2023)
Article
Medicine, General & Internal
Xiaofang Zhang, Xiaomin Liu, Bin Zhang, Jie Dong, Bin Zhang, Shujun Zhao, Suxiao Li
Summary: This study proposes an accurate segmentation method based on an improved U-Net convolutional network for different types of lung nodules on computed tomography images. The method includes two phases: segmenting lung parenchyma and correcting the lung contour, and extracting image patches of lung nodules with corresponding ground truth to train the network. Results show that the segmentation performance of Dice loss is superior to mean square error and Binary_crossentropy loss, and the alpha-hull algorithm and batch normalization can effectively improve segmentation performance.
Article
Engineering, Biomedical
Qi Ge, Tienan Xia, Yan Qiu, Jinxin Liu, Guanning Shang, Bin Liu
Summary: A semiautomatic segmentation method for pelvic bone tumors based on CT-MR multimodal images is proposed in this study. The method combines multiple medical prior knowledge and image segmentation algorithms. The results show that the proposed algorithm can accurately segment bone tumors in pelvic MR images and provide assistance for preservation surgery.
INTERNATIONAL JOURNAL FOR NUMERICAL METHODS IN BIOMEDICAL ENGINEERING
(2023)
