Article
Clinical Neurology
Sreevani Katabathula, Qinyong Wang, Rong Xu
Summary: The study introduced a new deep learning model DenseCNN2 for AD classification by combining hippocampus segmentations and global shape features, which outperformed other deep learning models in terms of accuracy and performance.
ALZHEIMERS RESEARCH & THERAPY
(2021)
Article
Computer Science, Artificial Intelligence
Ruizhi Han, Zhulin Liu, C. L. Philip Chen
Summary: This paper presents a new variant model of the Broad Learning System (BLS) for accurate diagnosis of Alzheimer's disease (AD) and Mild Cognitive Impairment (MCI) using MRI images. The proposed model integrates multi-scale convolution features and abstract features to achieve precise diagnosis. Experimental results demonstrate that the proposed model outperforms other methods in AD and MCI diagnostic tasks.
APPLIED SOFT COMPUTING
(2022)
Article
Biotechnology & Applied Microbiology
T. Illakiya, Karthik Ramamurthy, M. V. Siddharth, Rashmi Mishra, Ashish Udainiya
Summary: Alzheimer's disease is a progressive neurological problem that affects memory and thinking skills. Accurate detection is challenging, but the proposed Adaptive Hybrid Attention Network (AHANet) with Enhanced Non-Local Attention (ENLA) and Coordinate Attention modules shows promising results, extracting global and local features from MRI images. The network also incorporates an Adaptive Feature Aggregation (AFA) module to effectively fuse global and local features. The proposed network achieved a classification accuracy of 98.53% on the ADNI dataset.
BIOENGINEERING-BASEL
(2023)
Article
Mathematical & Computational Biology
Xiaoxiao Chen, Linghui Li, Ashutosh Sharma, Gaurav Dhiman, S. Vimal
Summary: Alzheimer's disease is an irrepressible neurological brain disorder, and early detection and proper treatment are crucial for preventing brain tissue damage. The study used a CNN model to explore the segmentation effects on MR imaging for Alzheimer's diagnosis and nursing, finding that the CNN model demonstrated higher segmentation precision.
INTERDISCIPLINARY SCIENCES-COMPUTATIONAL LIFE SCIENCES
(2022)
Article
Computer Science, Theory & Methods
Bin Lu, Hui-Xian Li, Zhi-Kai Chang, Le Li, Ning-Xuan Chen, Zhi-Chen Zhu, Hui-Xia Zhou, Xue-Ying Li, Yu-Wei Wang, Shi-Xian Cui, Zhao-Yu Deng, Zhen Fan, Hong Yang, Xiao Chen, Paul M. Thompson, Francisco Xavier Castellanos, Chao-Gan Yan
Summary: This study constructed a practical AD diagnostic classifier based on brain MRI using deep learning and transfer learning, achieving high accuracy in AD diagnosis on multiple independent datasets through training with a large-scale dataset, and making good predictions for MCI patients converting to AD.
JOURNAL OF BIG DATA
(2022)
Article
Biotechnology & Applied Microbiology
Pauline Shan Qing Yeoh, Khin Wee Lai, Siew Li Goh, Khairunnisa Hasikin, Xiang Wu, Pei Li
Summary: This study investigates the feasibility of using well-known convolutional neural network (CNN) structures (ResNet, DenseNet, VGG, and AlexNet) to distinguish knees with and without osteoarthritis (OA). The results show that 3D convolutional neural networks using 3D convolutional layers have potential in knee osteoarthritis diagnosis. Transfer learning by transforming 2D pre-trained weights into 3D enhances the performance of the models. This study suggests the possibility of clinical diagnostic aid for knee osteoarthritis using 3DCNN.
FRONTIERS IN BIOENGINEERING AND BIOTECHNOLOGY
(2023)
Article
Multidisciplinary Sciences
Jie Dong, Shujun Zhao, Yun Meng, Yong Zhang, Suxiao Li
Summary: This study aimed to explore the application value of a magnetic resonance imaging (MRI) image reconstruction model based on complex convolutional neural network (CCNN) in diagnosing and predicting the prognosis of cerebral infarction. Two new MRI image reconstruction models, D-FDRN and D-IDRN, were created by integrating two image reconstruction methods based on the CCNN algorithm. The study found that the PSNR and SSIM values of the MRI reconstructed image using the D-IDRN algorithm were higher than those of other algorithms, and analyzed the correlation between vein abnormality grading (VABG) and infarct size as well as the degree of stenosis of the responsible vessel. The changes in ADC value and DCavg value in the central area of the infarct could be used for diagnosing cerebral infarction.
Review
Computer Science, Interdisciplinary Applications
T. Illakiya, R. Karthik
Summary: Deep learning algorithms have a significant impact on medical image processing in the field of research, providing vital aid for radiologists in accurate disease diagnosis. This study aims to emphasize the importance of deep learning models in detecting Alzheimer's Disease (AD) and analyzes various methods used for AD detection. 103 research articles were analyzed, focusing on deep learning techniques such as Convolutional Neural Networks (CNNs), Recurrent Neural Networks (RNNs), and Transfer Learning (TL). The review examines the use of neuroimaging modalities like Positron Emission Tomography (PET), Magnetic Resonance Imaging (MRI), etc. in deep learning methods for AD detection.
Article
Chemistry, Analytical
Robin Cabeza-Ruiz, Luis Velazquez-Perez, Alejandro Linares-Barranco, Roberto Perez-Rodriguez
Summary: The study proposes the use of convolutional neural networks to automatically segment cerebellar fissures from brain magnetic resonance imaging. Three models based on the same CNN architecture are presented, achieving good performance in terms of precision and efficiency.
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
Computer Science, Information Systems
Aya Gamal, Mustafa Elattar, Sahar Selim
Summary: This study proposes a new method for early detection of Alzheimer's disease using a computer-aided system. By processing MRI images and conducting multiple experiments, an ensemble learning approach is introduced, which outperforms previous studies in distinguishing different disease stages and multi-class tasks.
Review
Mathematical & Computational Biology
Zhen Zhao, Joon Huang Chuah, Khin Wee Lai, Chee-Onn Chow, Munkhjargal Gochoo, Samiappan Dhanalakshmi, Na Wang, Wei Bao, Xiang Wu
Summary: Alzheimer's disease (AD) is a neurodegenerative disorder that affects memory and cognitive function in elderly individuals. While there is no cure for AD, early detection is crucial for slowing down the disease progression. Recent studies have shown that deep learning approaches have achieved success in AD diagnosis using Magnetic Resonance Imaging (MRI). This paper reviews conventional machine learning methods such as support vector machine (SVM), random forest (RF), convolutional neural network (CNN), autoencoder, deep learning, and transformer, as well as commonly used feature extractors and input forms for CNN. It also discusses challenges, trade-offs, and suggestions for preprocessing techniques, deep learning, conventional machine learning methods, new techniques, and input type selection.
FRONTIERS IN COMPUTATIONAL NEUROSCIENCE
(2023)
Article
Computer Science, Information Systems
Fazal Ur Rehman Faisal, Goo-Rak Kwon
Summary: This study aims to develop a deep learning method for extracting valuable Alzheimer's disease biomarkers from structural magnetic resonance imaging (sMRI) and classifying brain images into different groups. The proposed method shows superior results compared to existing methods, with reduced parameters and computation complexity.
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
Medicine, General & Internal
Noman Raza, Asma Naseer, Maria Tamoor, Kashif Zafar
Summary: Alzheimer's disease is a slow neurological disorder that affects the thought process and consciousness of individuals, particularly elderly people above 60 years of age. This research focuses on segmenting and classifying Alzheimer's disease Magnetic Resonance Imaging (MRI) using transfer learning and customizing a convolutional neural network (CNN) trained on brain gray matter (GM) segmented images. By using a pre-trained deep learning model and applying transfer learning, the proposed model achieves an overall accuracy of 97.84% when tested over 10, 25, and 50 epochs.
Article
Clinical Neurology
Lydia Chougar, Francois-Xavier Lejeune, Johann Faouzi, Benjamin Morino, Alice Faucher, Nadine Hoyek, David Grabli, Florence Cormier, Marie Vidailhet, Jean-Christophe Corvol, Olivier Colliot, Bertrand Degos, Stephane Lehericy
Summary: This study evaluates the value of combining clinically feasible manual measurements and morphometric measurements to improve the discrimination of parkinsonian syndromes. The results show that combining R2* and MD measurements with morphometric biomarkers can better differentiate parkinsonian syndromes.
PARKINSONISM & RELATED DISORDERS
(2023)
Article
Computer Science, Artificial Intelligence
Simona Bottani, Ninon Burgos, Aurelien Maire, Dario Saracino, Sebastian Stroer, Didier Dormont, Olivier Colliot, Alzheimers Dis Neuroimaging Initiat, APPRIMAGE Study Grp
Summary: This study examines the performance of computer-aided diagnosis methods on clinical routine data and compares it to research data. The results indicate that the performance is strongly biased upward due to confounding factors like image quality and contrast agent injection, and overall much lower than on research data.
MEDICAL IMAGE ANALYSIS
(2023)
Editorial Material
Anesthesiology
Theodore Soulier, Olivier Colliot, Nicholas Ayache, Benjamin Rohaut
ANAESTHESIA CRITICAL CARE & PAIN MEDICINE
(2023)
Article
Anatomy & Morphology
Kevin de Matos, Claire Cury, Lydia T. Chougar, Lachlan Strike, Thibault Rolland, Maximilien Riche, Lisa Hemforth, Alexandre Martin, Tobias Banaschewski, Arun L. W. Bokde, Sylvane Desrivieres, Herta Flor, Antoine Grigis, Hugh Garavan, Penny Gowland, Andreas Heinz, Rudiger Bruhl, Jean-Luc Martinot, Marie-Laure Paillere Martinot, Eric Artiges, Frauke Nees, Dimitri Papadopoulos Orfanos, Herve Lemaitre, Tomas Paus, Luise Poustka, Sarah Hohmann, Sabina H. Millenet, Juliane N. Frohner, Michael Smolka, Nilakshi Vaidya, Henrik Walter, Robert Whelan, Gunter Schumann, Vincent Frouin, Meritxell Bach Cuadra, Olivier Colliot, Baptiste Couvy-Duchesne
Summary: The temporo-basal region of the human brain consists of the collateral, occipito-temporal, and rhinal sulci. In this study, we manually evaluated the connections between these sulci using MRI data from nearly 3400 individuals, including twins. We found hemisphere-dependent frequency and sexual dimorphism in these connections, with differences between males and females.
BRAIN STRUCTURE & FUNCTION
(2023)
Proceedings Paper
Computer Science, Artificial Intelligence
Elina Thibeau-Sutre, Jelmer M. Wolterink, Olivier Colliot, Ninon Burgos
Summary: This study aims to investigate whether attribution maps obtained using gradient back-propagation can correctly highlight the patterns of disease subtypes discovered by a deep learning classifier. Due to the difficulty in directly evaluating the accuracy of attribution maps on medical images, we used synthetic data mimicking the difference between brain MRI of controls and demented patients to design more reliable evaluation criteria. The results showed that attribution maps may mix the regions associated with different subtypes for small data sets, while accurately characterizing both subtypes using a large data set. We also proposed simple data augmentation techniques and demonstrated that they can improve the coherence of explanations for a small data set.
MEDICAL IMAGING 2023
(2023)
Proceedings Paper
Computer Science, Artificial Intelligence
Guanghui Fu, Rosana El Jurdi, Lydia Chougar, Didier Dormont, Romain Valabregue, Stephane Lehericy, Olivier Colliot
Summary: Deep learning methods have achieved impressive results in 3D medical image segmentation. However, when guided by voxel-level information alone, the resulting segmentations may contain anatomically aberrant structures. To address this issue, a novel loss function is proposed in this paper to introduce topological priors in deep learning-based segmentation. The proposed method is computationally efficient and easy to implement.
MEDICAL IMAGING 2023
(2023)
Proceedings Paper
Computer Science, Artificial Intelligence
Ravi Hassanaly, Simona Bottani, Benoit Sauty, Olivier Colliot, Ninon Burgos
Summary: Unsupervised anomaly detection using deep learning models is a popular approach for computer-aided diagnosis, especially in neuroimaging. In this work, the focus is on detecting anomalies from FDG PET images of patients with Alzheimer's disease, which can be subtle and difficult to evaluate. To address this, the researchers propose a framework for evaluating unsupervised anomaly detection approaches by simulating realistic anomalies from healthy images. They demonstrate the use of this framework by evaluating an approach based on a 3D variational autoencoder.
MEDICAL IMAGING 2023
(2023)
Proceedings Paper
Computer Science, Artificial Intelligence
Sophie Loizillon, Simona Bottani, Aurelien Maire, Sebastian Stroer, Didier Dormont, Olivier Colliot, Ninon Burgos
Summary: Clinical data warehouses (CDWs) provide an opportunity for developing computational tools by containing the medical data of millions of patients. This paper proposes a CNN for the automatic detection of motion in 3D T1-weighted brain MRI to fully exploit CDWs. The framework achieved excellent accuracy in excluding images with severe motion, but weaker performance in detecting mild motion artefacts compared to human raters.
MEDICAL IMAGING 2023
(2023)
Article
Clinical Neurology
William Coath, Marc Modat, M. Jorge J. Cardoso, Pawel J. A. Markiewicz, Christopher A. D. Lane, Thomas D. Parker, Ashvini M. Keshavan, Sarah M. E. Buchanan, Sarah E. J. Keuss, Matthew J. Harris, Ninon Burgos, John Dickson, Anna L. Barnes, David L. Thomas, Daniel B. Beasley, Ian B. Malone, Andrew Wong, Kjell A. Erlandsson, Benjamin A. Thomas, Michael Scholl, Sebastien Ourselin, Marcus C. Richards, Nick C. M. Fox, Jonathan M. M. Schott, David M. Cash
Summary: The Centiloid scale aims to harmonize Aβ PET measures across different analysis methods. This study investigated the Centiloid transformation with PET/MRI data, finding that the transformation is valid but further understanding of the effects of acquisition or biological factors on using white matter as a reference is needed.
ALZHEIMER'S & DEMENTIA: DIAGNOSIS, ASSESSMENT & DISEASE MONITORING
(2023)
Article
Neuroimaging
Arya Yazdan-Panah, Marius Schmidt-Mengin, Vito A. G. Ricigliano, Theodore Soulier, Bruno Stankoff, Olivier Colliot
Summary: Choroid Plexuses (ChP) play a crucial role in producing cerebrospinal fluid (CSF) and have been found to undergo volumetric changes in various neurological diseases. To investigate their role, an automated and reliable ChP segmentation tool is needed for large-scale studies. In this study, we propose a novel 2-step 3D U-Net-based automatic method for ChP segmentation, which minimizes preprocessing steps for ease of use and lower memory requirements.
NEUROIMAGE-CLINICAL
(2023)
Article
Clinical Neurology
Cecile Di Folco, Raphael Couronne, Isabelle Arnulf, Graziella Mangone, Smaranda Leu-Semenescu, Pauline Dodet, Marie Vidailhet, Jean-Christophe Corvol, Stephane Lehericy, Stanley Durrleman
Summary: This study proposes a disease course map for Parkinson's disease (PD) and investigates the progression profiles of patients with or without rapid eye movement sleep behavioral disorders (RBD). The findings reveal distinct patterns of progression between PD patients with and without RBD, emphasizing the importance of understanding heterogeneity in PD progression for precision medicine.
MOVEMENT DISORDERS
(2023)
Article
Computer Science, Artificial Intelligence
Clement Chadebec, Elina Thibeau-Sutre, Ninon Burgos, Stephanie Allassonniere
Summary: This paper proposes a new method to perform reliable data augmentation in the High Dimensional Low Sample Size (HDLSS) setting using a geometry-based variational autoencoder (VAE). The method demonstrates robustness to datasets, classifiers, and training sample sizes, and is validated in a medical imaging classification task, showing significant improvement in classification metrics.
IEEE TRANSACTIONS ON PATTERN ANALYSIS AND MACHINE INTELLIGENCE
(2023)
Article
Computer Science, Artificial Intelligence
Hong Liu, Dong Wei, Donghuan Lu, Xiaoying Tang, Liansheng Wang, Yefeng Zheng
Summary: This study proposes a framework based on hybrid 2D-3D convolutional neural networks for obtaining continuous 3D retinal layer surfaces from OCT volumes. The framework works well with both full and sparse annotations and utilizes alignment displacement vectors and layer segmentation to align the B-scans and segment the layers. Experimental results show that the framework outperforms state-of-the-art 2D deep learning methods in terms of layer segmentation accuracy and cross-B-scan 3D continuity.
MEDICAL IMAGE ANALYSIS
(2024)
Article
Computer Science, Artificial Intelligence
Simon Oxenford, Ana Sofia Rios, Barbara Hollunder, Clemens Neudorfer, Alexandre Boutet, Gavin J. B. Elias, Jurgen Germann, Aaron Loh, Wissam Deeb, Bryan Salvato, Leonardo Almeida, Kelly D. Foote, Robert Amaral, Paul B. Rosenberg, David F. Tang-Wai, David A. Wolk, Anna D. Burke, Marwan N. Sabbagh, Stephen Salloway, M. Mallar Chakravarty, Gwenn S. Smith, Constantine G. Lyketsos, Michael S. Okun, William S., Zoltan Mari, Francisco A. Ponce, Andres Lozano, Wolf-Julian Neumann, Bassam Al-Fatly, Andreas Horn
Summary: Spatial normalization is a method to map subject brain images to an average template brain, allowing comparison of brain imaging results. We introduce a novel tool called WarpDrive, which enables manual refinements of image alignment after automated registration. The tool improves accuracy of data representation and aids in understanding patient outcomes.
MEDICAL IMAGE ANALYSIS
(2024)
Article
Computer Science, Artificial Intelligence
Ricards Marcinkevics, Patricia Reis Wolfertstetter, Ugne Klimiene, Kieran Chin-Cheong, Alyssia Paschke, Julia Zerres, Markus Denzinger, David Niederberger, Sven Wellmann, Ece Ozkan, Christian Knorr, Julia E. Vogt
Summary: This study presents interpretable machine learning models for predicting the diagnosis, management, and severity of suspected appendicitis using ultrasound images. The proposed models utilize concept bottleneck models (CBM) that facilitate interpretation and intervention by clinicians, without compromising performance or requiring time-consuming image annotation.
MEDICAL IMAGE ANALYSIS
(2024)
Article
Computer Science, Artificial Intelligence
Jian-Qing Zheng, Ziyang Wang, Baoru Huang, Ngee Han Lim, Bartlomiej W. Papiez
Summary: This article introduces a new method for medical image registration, which utilizes a separable motion backbone and a residual aligner module to better handle the discontinuous motion of multiple neighboring objects. The proposed method achieves excellent registration results on abdominal CT scans and lung CT scans.
MEDICAL IMAGE ANALYSIS
(2024)
Article
Computer Science, Artificial Intelligence
Xiangqiong Wu, Guanghua Tan, Hongxia Luo, Zhilun Chen, Bin Pu, Shengli Li, Kenli Li
Summary: This study develops a user-friendly framework for the automated diagnosis of thyroid nodules in ultrasound videos, simulating the diagnostic workflow of radiologists. By interpreting image characteristics and modeling temporal contextual information, the efficiency and generalizability of the diagnosis can be improved.
MEDICAL IMAGE ANALYSIS
(2024)
Article
Computer Science, Artificial Intelligence
Riddhish Bhalodia, Shireen Elhabian, Jadie Adams, Wenzheng Tao, Ladislav Kavan, Ross Whitaker
Summary: This paper introduces DeepSSM, a deep learning-based framework for image-to-shape modeling. By learning the functional mapping from images to low-dimensional shape descriptors, DeepSSM can directly infer statistical representation of anatomy from 3D images. Compared to traditional methods, DeepSSM eliminates the need for heavy manual preprocessing and segmentation, and significantly improves computational time.
MEDICAL IMAGE ANALYSIS
(2024)
Article
Computer Science, Artificial Intelligence
Florentin Liebmann, Marco von Atzigen, Dominik Stutz, Julian Wolf, Lukas Zingg, Daniel Suter, Nicola A. Cavalcanti, Laura Leoty, Hooman Esfandiari, Jess G. Snedeker, Martin R. Oswald, Marc Pollefeys, Mazda Farshad, Philipp Furnstahl
Summary: This study presents a marker-less approach for automatic registration and real-time navigation of lumbar spinal fusion surgery using a deep neural network, avoiding radiation exposure and surgical errors. The method was validated on an ex-vivo surgery and a public dataset.
MEDICAL IMAGE ANALYSIS
(2024)
Article
Computer Science, Artificial Intelligence
Piyush Tiwary, Kinjawl Bhattacharyya, A. P. Prathosh
Summary: Domain shift refers to the change of distributional characteristics between training and testing datasets, leading to performance drop. For medical image tasks, domain shift can be caused by changes in imaging modalities, devices, and staining mechanisms. Existing approaches based on generative models suffer from training difficulties and lack of diversity. In this paper, the authors propose the use of energy-based models (EBMs) for unpaired image-to-image translation in medical images. The proposed method, called Cycle Consistent Twin EBMs (CCT-EBM), employs a pair of EBMs in the latent space of an Auto-Encoder to ensure translation symmetry and coupling between domains.
MEDICAL IMAGE ANALYSIS
(2024)
Article
Computer Science, Artificial Intelligence
Yutong Xie, Jianpeng Zhang, Lingqiao Liu, Hu Wang, Yiwen Ye, Johan Verjans, Yong Xia
Summary: This paper proposes a hybrid pre-training paradigm that combines self-supervised learning and supervised learning to improve the representation quality for medical image segmentation tasks. It introduces a reference task in self-supervised learning and optimizes the model using a gradient matching method. The experimental results demonstrate the effectiveness of this approach on multiple medical image segmentation benchmarks.
MEDICAL IMAGE ANALYSIS
(2024)
Article
Computer Science, Artificial Intelligence
Youyi Song, Jing Zou, Kup-Sze Choi, Baiying Lei, Jing Qin
Summary: Cell classification is crucial for intelligent cervical cancer screening, but the variation in cells' appearance and shape poses challenges. A new learning algorithm, worse-case boosting, is proposed to improve classification accuracy for under-represented data. Experimental results demonstrate the effectiveness of this algorithm in two publicly available datasets, achieving a 4% improvement in accuracy.
MEDICAL IMAGE ANALYSIS
(2024)
Article
Computer Science, Artificial Intelligence
Sangjoon Park, Eun Sun Lee, Kyung Sook Shin, Jeong Eun Lee, Jong Chul Ye
Summary: The increasing demand for AI systems to monitor human errors and abnormalities in healthcare presents challenges. This study presents a model called Medical X-VL, which is tailored for the medical domain and outperformed current state-of-the-art models in two medical image datasets. The model enables various zero-shot tasks for monitoring AI in the medical domain.
MEDICAL IMAGE ANALYSIS
(2024)
Article
Computer Science, Artificial Intelligence
Anna Klimovskaia Susmelj, Berkan Lafci, Firat Ozdemir, Neda Davoudi, Xose Luis Dean-Ben, Fernando Perez-Cruz, Daniel Razansky
Summary: Optoacoustic imaging is a technique that uses optical excitation and ultrasound detection for biological tissue imaging. The quality of the images depends on the extent of tomographic coverage provided by the ultrasound detector arrays. However, full coverage is not always possible due to experimental constraints. The proposed signal domain adaptation network aims to reduce limited-view artifacts in the images.
MEDICAL IMAGE ANALYSIS
(2024)
Article
Computer Science, Artificial Intelligence
Srijay Deshpande, Muhammad Dawood, Fayyaz Minhas, Nasir Rajpoot
Summary: In this work, a novel framework called SynCLay is proposed for automated synthesis of histology images based on user-defined cellular layouts. The framework can generate realistic and high-quality histology images with different cellular arrangements, which is helpful for studying the role of cells in the tumor microenvironment. The framework integrates a nuclear segmentation and classification model to refine nuclear structures and generate nuclear masks. Evaluation using quantitative metrics and feedback from pathologists shows that the synthetic images generated by SynCLay have high realism scores and can accurately differentiate between benign and malignant tumors.
MEDICAL IMAGE ANALYSIS
(2024)
Article
Computer Science, Artificial Intelligence
Ahmed H. Shahin, An Zhao, Alexander C. Whitehead, Daniel C. Alexander, Joseph Jacob, David Barber
Summary: Survival analysis is a valuable tool in healthcare for predicting the time to specific events. This paper introduces CenTime, a novel approach that directly estimates the time to event. The method performs well with censored data and can be easily integrated with deep learning models. Compared to standard methods, CenTime offers superior performance in predicting event time while maintaining comparable ranking performance.
MEDICAL IMAGE ANALYSIS
(2024)
Article
Computer Science, Artificial Intelligence
Bingyuan Liu, Jose Dolz, Adrian Galdran, Riadh Kobbi, Ismail Ben Ayed
Summary: Most segmentation losses, such as CE and Dice, are variants of the Cross-Entropy or Dice losses. This work provides a theoretical analysis that shows a deeper connection between CE and Dice than previously thought. From a constrained-optimization perspective, both CE and Dice decompose into similar ground-truth matching terms and region-size penalty terms. The analysis uncovers hidden region-size biases: Dice has an intrinsic bias towards extremely imbalanced solutions, while CE implicitly encourages the ground-truth region proportions. Based on this analysis, a principled and simple solution is proposed to explicitly control the region-size bias.
MEDICAL IMAGE ANALYSIS
(2024)