In this paper, we propose the K-Shot Contrastive Learning (KSCL) of visual features by applying multiple augmentations to investigate the sample variations within individual instances. It aims to combine the advantages of inter-instance discrimination by learning discriminative features to distinguish between different instances, as well as intra-instance variations by matching queries against the variants of augmented samples over instances. Particularly, for each instance, it constructs an instance subspace to model the configuration of how the significant factors of variations in K-shot augmentations can be combined to form the variants of augmentations. Given a query, the most relevant variant of instances is then retrieved by projecting the query onto their subspaces to predict the positive instance class. This generalizes the existing contrastive learning that can be viewed as a special one-shot case. An eigenvalue decomposition is performed to configure instance subspaces, and the embedding network can be trained end-to-end through the differentiable subspace configuration. Experiment results demonstrate the proposed K-shot contrastive learning achieves superior performances to the state-of-the-art unsupervised methods.
Download full-text PDF |
Source |
---|---|
http://dx.doi.org/10.1109/TPAMI.2021.3082567 | DOI Listing |
Nat Commun
January 2025
Department of Computational Biology and Medical Science, Graduate School of Frontier Sciences, the University of Tokyo, Tokyo, Japan.
Spatial transcriptomics is an essential application for investigating cellular structures and interactions and requires multimodal information to precisely study spatial domains. Here, we propose STAIG, a deep-learning model that integrates gene expression, spatial coordinates, and histological images using graph-contrastive learning coupled with high-performance feature extraction. STAIG can integrate tissue slices without prealignment and remove batch effects.
View Article and Find Full Text PDFAcad Radiol
January 2025
Medical Image Processing Group, 602 Goddard building, 3710 Hamilton Walk, Department of Radiology, University of Pennsylvania, Philadelphia, PA 19104 (M.L., M.A., J.K.U., Y.T., C.W., N.P., S.M., D.A.T.). Electronic address:
Rationale And Objectives: Cardiovascular toxicity is a well-known complication of thoracic radiation therapy (RT), leading to increased morbidity and mortality, but existing techniques to predict cardiovascular toxicity have limitations. Predictive biomarkers of cardiovascular toxicity may help to maximize patient outcomes.
Methods: The machine learning optimal biomarker (OBM) method was employed to predict development of cardiotoxicity (based on serial echocardiographic measurements of left ventricular ejection fraction and longitudinal strain) from computed tomography (CT) images in patients with thoracic malignancy undergoing RT.
Phys Med Biol
January 2025
School of Biomedical Engineering, ShanghaiTech University, No. 1 Zhongke Road, Pudong New Area, Shanghai, Shanghai, 201210, CHINA.
Objective: This study aims to propose a dual-domain network that not only reduces scatter artifacts but also retains structure details in CBCT.
Approach: The proposed network comprises a projection-domain sub-network and an image-domain sub-network. The projection-domain sub-network utilizes a division residual network to amplify the difference between scatter signals and imaging signals, facilitating the learning of scatter signals.
Comput Biol Med
January 2025
School of Automation Science and Engineering, South China University of Technology, Guangzhou, China. Electronic address:
Breast cancer poses a significant health threat worldwide. Contrastive learning has emerged as an effective method to extract critical lesion features from mammograms, thereby offering a potent tool for breast cancer screening and analysis. A crucial aspect of contrastive learning is negative sampling, where the selection of hard negative samples is essential for driving representations to retain detailed lesion information.
View Article and Find Full Text PDFJ Imaging Inform Med
January 2025
Department of Radiation Oncology, Henry Ford Health, Detroit, MI, USA.
Automatic segmentation of angiographic structures can aid in assessing vascular disease. While recent deep learning models promise automation, they lack validation on interventional angiographic data. This study investigates the feasibility of angiographic segmentation using in-context learning with the UniverSeg model, which is a cross-learning segmentation model that lacks inherent angiographic training.
View Article and Find Full Text PDFEnter search terms and have AI summaries delivered each week - change queries or unsubscribe any time!