Category Ranking

98%

Total Visits

921

Avg Visit Duration

2 minutes

Citations

20

Article Abstract

Deep learning (DL) requires large amounts of labeled data, which is extremely time-consuming and laborintensive to obtain for medical image segmentation tasks. Metalearning focuses on developing learning strategies that enable quick adaptation to new tasks with limited labeled data. However, rich-class medical image segmentation datasets for constructing meta-learning multi-tasks are currently unavailable. In addition, data collected from various healthcare sites and devices may present significant distribution differences, potentially degrading model's performance. In this paper, we propose a task augmentation-based meta-learning method for retinal image segmentation (TAMS) to meet labor-intensive annotation demand. A retinal Lesion Simulation Algorithm (LSA) is proposed to automatically generate multi-class retinal disease datasets with pixel-level segmentation labels, such that metalearning tasks can be augmented without collecting data from various sources. In addition, a novel simulation function library is designed to control generation process and ensure interpretability. Moreover, a generative simulation network (GSNet) with an improved adversarial training strategy is introduced to maintain high-quality representations of complex retinal diseases. TAMS is evaluated on three different OCT and CFP image datasets, and comprehensive experiments have demonstrated that TAMS achieves superior segmentation performance than state-of-the-art models.

Download full-text PDF

Source
http://dx.doi.org/10.1109/TPAMI.2025.3579271DOI Listing

Publication Analysis

Top Keywords

image segmentation
12
task augmentation-based
8
augmentation-based meta-learning
8
labeled data
8
medical image
8
segmentation
6
meta-learning segmentation
4
segmentation method
4
method retinopathy
4
retinopathy deep
4

Similar Publications

Purpose To assess the effectiveness of an explainable deep learning (DL) model, developed using multiparametric MRI (mpMRI) features, in improving diagnostic accuracy and efficiency of radiologists for classification of focal liver lesions (FLLs). Materials and Methods FLLs ≥ 1 cm in diameter at mpMRI were included in the study. nn-Unet and Liver Imaging Feature Transformer (LIFT) models were developed using retrospective data from one hospital (January 2018-August 2023).

View Article and Find Full Text PDF

Patients with inflammation-associated coronary artery disease (CAD) may exhibit rapid progression and require regular coronary imaging. To evaluate the diagnostic performance of spectral photon-counting detector (PCD) coronary CTA with reduced radiation and contrast media doses for detecting coronary stenosis and in-stent restenosis in patients with inflammation-associated CAD. This prospective study enrolled patients with inflammation-associated CAD from January 2023 to March 2024.

View Article and Find Full Text PDF

Parasagittal dural space and arachnoid granulations morphology in pre-clinical and early clinical multiple sclerosis.

Mult Scler

September 2025

Neuroimaging Unit, Neuroimmunology Division, Department of Neurology, Vanderbilt University Medical Center, Nashville, TN, USA; Department of Neurology, VA Medical Center, TN Valley Healthcare System, Nashville, TN, USA.

Background: There is limited knowledge on the post-glymphatic structures such as the parasagittal dural (PSD) space and the arachnoid granulations (AGs) in multiple sclerosis (MS).

Objectives: To evaluate differences in volume and macromolecular content of PSD and AG between people with newly diagnosed MS (pwMS), clinically isolated syndrome (pwCIS), or radiologically isolated syndrome (pwRIS) and healthy controls (HCs) and their associations with clinical and radiological disease measures.

Methods: A total of 69 pwMS, pwCIS, pwRIS, and HCs underwent a 3.

View Article and Find Full Text PDF

Purpose: To evaluate the impact of an optimized online adaptive radiation therapy workflow on physician involvement.

Methods And Materials: Data from a prospective phase 2 trial involving 34 prostate cancer patients treated with cone beam computed tomography (CBCT)-based online adaptive radiation therapy (62 Gy in 20 fractions) were analyzed. Manual interventions were required for 2 steps in the workflow: radiation therapy technologist review and adjustment of automatically segmented organs, guiding target segmentation, so-called "influencer," while physicians reviewed and refined the targets.

View Article and Find Full Text PDF