paper-with-me

Papers

Test-time augmentation-based active learning and self-training for label-efficient segmentation

2023-08-21 · Bella Specktor-Fadida, Anna Levchakov, Dana Schonberger, Liat Ben-Sira, Dafna Ben-Bashat, Leo Joskowicz

Deep learning techniques depend on large datasets whose annotation is time-consuming. To reduce annotation burden, the self-training (ST) and active-learning (AL) methods have been developed as well as methods that combine them in an iterative fashion. However, it remains unclear when each method is the most useful, and when it is advantageous to combine them. In this paper, we propose a new method that combines ST with AL using Test-Time Augmentations (TTA). First, TTA is performed on an initial teacher network. Then, cases for annotation are selected based on the lowest estimated Dice score. Cases with high estimated scores are used as soft pseudo-labels for ST. The selected annotated cases are trained with existing annotated cases and ST cases with border slices annotations. We demonstrate the method on MRI fetal body and placenta segmentation tasks with different data variability characteristics. Our results indicate that ST is highly effective for both tasks, boosting performance for in-distribution (ID) and out-of-distribution (OOD) data. However, while self-training improved the performance of single-sequence fetal body segmentation when combined with AL, it slightly deteriorated performance of multi-sequence placenta segmentation on ID data. AL was helpful for the high variability placenta data, but did not improve upon random selection for the single-sequence body data. For fetal body segmentation sequence transfer, combining AL with ST following ST iteration yielded a Dice of 0.961 with only 6 original scans and 2 new sequence scans. Results using only 15 high-variability placenta cases were similar to those using 50 cases. Code is available at: https://github.com/Bella31/TTA-quality-estimation-ST-AL

📄 PDF Abstract BibTeX arXiv:2308.10727

Code (0)

등록된 구현이 없습니다.

Tasks

Active LearningPlacenta SegmentationSegmentation

Similar Papers 제목 키워드 기반

End-to-End Augmentation Hyperparameter Tuning for Self-Supervised Anomaly Detection

2023-06-21 · Jaemin Yoo, Lingxiao Zhao, Leman Akoglu

Self-supervised learning (SSL) has emerged as a promising paradigm that presents self-generated supervisory signals to real-world problems, bypassing the extensive manual labeling burden. SSL is especially attractive for…

Anomaly DetectionData AugmentationSelf-Supervised Anomaly DetectionSelf-Supervised Learning+1

The University of Arizona at SemEval-2021 Task 10: Applying Self-training, Active Learning and Data Augmentation to Source-free Domain Adaptation

2021-08-01 · SEMEVAL 2021 · Xin Su, Yiyun Zhao, Steven Bethard

This paper describes our systems for negation detection and time expression recognition in SemEval 2021 Task 10, Source-Free Domain Adaptation for Semantic Processing. We show that self-training, active learning and data…

Active LearningData AugmentationDomain AdaptationNegation+2

Learning Loss for Test-Time Augmentation

2020-10-22 · NeurIPS 2020 12 · Ildoo Kim, Younghoon Kim, Sungwoong Kim

Data augmentation has been actively studied for robust neural networks. Most of the recent data augmentation methods focus on augmenting datasets during the training phase. At the testing phase, simple transformations ar…

Data Augmentationimage-classificationImage Classification

Improved automated lesion segmentation in whole-body FDG/PET-CT via Test-Time Augmentation

2022-10-14 · Sepideh Amiri, Bulat Ibragimov

Numerous oncology indications have extensively quantified metabolically active tumors using positron emission tomography (PET) and computed tomography (CT). F-fluorodeoxyglucose-positron emission tomography (FDG-PET) is …

Computed Tomography (CT)Lesion SegmentationSegmentationTumor Segmentation

TeSLA: Test-Time Self-Learning With Automatic Adversarial Augmentation

2023-03-17 · CVPR 2023 1 · Devavrat Tomar, Guillaume Vray, Behzad Bozorgtabar, Jean-Philippe Thiran

Most recent test-time adaptation methods focus on only classification tasks, use specialized network architectures, destroy model calibration or rely on lightweight information from the source domain. To tackle these iss…

Knowledge DistillationSelf-LearningTest-time Adaptation