Double-Uncertainty Weighted Method for Semi-supervised Learning

Wang, Yixin; Zhang, Yao; Tian, Jiang; Zhong, Cheng; Shi, Zhongchao; Zhang, Yang; He, Zhiqiang

doi:10.1007/978-3-030-59710-8_53

Cited by 81 publications

(47 citation statements)

References 7 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Tabel Table 4 summarizes the quantitative results of our proposed method and several state-of-the-art methods, including LG-ER-MT [ 27 ], DUWM [ 28 ], MC-Net [ 29 ], V-net [ 30 ], Bayesian V-net, and AJSQnet [ 31 ]. Among them, LG-ER-MT, DUWM, and MC-Net utilized the semi-supervised strategy with uncertainty prediction, while V-net, Bayesian V-net, and AJSQnet are trained by all labeled data, and Bayesian V-Net utilized the Bayesian network to adapt the vanilla V-Net.…”

Section: Resultsmentioning

confidence: 99%

Two-Stage Segmentation Framework Based on Distance Transformation

Huang

Lin

Jiao

et al. 2021

Sensors

View full text Add to dashboard Cite

With the rise of deep learning, using deep learning to segment lesions and assist in diagnosis has become an effective means to promote clinical medical analysis. However, the partial volume effect of organ tissues leads to unclear and blurred edges of ROI in medical images, making it challenging to achieve high-accuracy segmentation of lesions or organs. In this paper, we assume that the distance map obtained by performing distance transformation on the ROI edge can be used as a weight map to make the network pay more attention to the learning of the ROI edge region. To this end, we design a novel framework to flexibly embed the distance map into the two-stage network to improve left atrium MRI segmentation performance. Furthermore, a series of distance map generation methods are proposed and studied to reasonably explore how to express the weight of assisting network learning. We conduct thorough experiments to verify the effectiveness of the proposed segmentation framework, and experimental results demonstrate that our hypothesis is feasible.

show abstract

Section: Resultsmentioning

confidence: 99%

Two-Stage Segmentation Framework Based on Distance Transformation

Huang

Lin

Jiao

et al. 2021

Sensors

View full text Add to dashboard Cite

show abstract

“…Uncertainties for image segmentation are derived from general considerations of the statistical model, from resampling training data sets in ensemble approaches [36], or from modifications like Monte Carlo dropout of the predictive procedure [37]. For semi-supervised learning, the uncertainty can be used to judge whether the model provides accurate and confident prediction, which can be leveraged to further exploit the unlabeled data and has been applied to many semi-supervised medical image segmentation tasks [14], [15]. Wang et al [38] found that Monte Carlo dropout perform better for uncertainty estimation.…”

Section: Uncertainty Estimationmentioning

confidence: 99%

“…To demonstrate the effectiveness of our method, a comprehensive comparison with existing methods is conducted on LA dataset. We evaluate our method with comparisons to several recent stateof-the-art semi-supervised segmentation methods, including transformation-consistent self-ensembline (TCSE) [34], mean teacher (MT) [13], uncertainty-aware mean teacher (UA-MT) [14], entropy minimization (Entropy Mini) [46], shape-aware semi-supervised segmentation (SASS) [17], double uncertainty weighted method (DUWM) [15], dual task consistency learning (DTC) [18] and uncertainty rectified pyramid consistency (URPC) [35]. To ensure a fair comparison, we used the same network backbone in these methods.…”

Section: Performance Of Using Different Percentages Of Labeled Datamentioning

confidence: 99%

“…Specifically, the mean teacher (MT) model [13] has achieved great success by enforcing the consistency of predictions from perturbed inputs between student and teacher models. Following [13], many consistency learning methods [14], [15], [16] have been proposed. Besides, another line of researches [17], [18], [19] focus on building task-level regularization by adding auxiliary task to leverage boundary-based surface mismatch.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Uncertainty-Guided Mutual Consistency Learning for Semi-Supervised Medical Image Segmentation

Zhang¹,

Jiao²,

Liao³

et al. 2021

Preprint

View full text Add to dashboard Cite

Medical image segmentation is a fundamental and critical step in many clinical approaches. Semi-supervised learning has been widely applied to medical image segmentation tasks since it alleviates the heavy burden of acquiring expertexamined annotations and takes the advantage of unlabeled data which is much easier to acquire. Although consistency learning has been proven to be an effective approach by enforcing an invariance of predictions under different distributions, existing approaches cannot make full use of region-level shape constraint and boundary-level distance information from unlabeled data. In this paper, we propose a novel uncertainty-guided mutual consistency learning framework to effectively exploit unlabeled data by integrating intra-task consistency learning from up-todate predictions for self-ensembling and cross-task consistency learning from task-level regularization to exploit geometric shape information. The framework is guided by the estimated segmentation uncertainty of models to select out relatively certain predictions for consistency learning, so as to effectively exploit more reliable information from unlabeled data. We extensively validate our proposed method on two publicly available benchmark datasets: Left Atrium Segmentation (LA) dataset and Brain Tumor Segmentation (BraTS) dataset. Experimental results demonstrate that our method achieves performance gains by leveraging unlabeled data and outperforms existing semisupervised segmentation methods.

show abstract

“…The first stage aims to generate high-quality pseudo labels, and the second stage aims to use pseudo labels to retrain the network to regularize features for both labeled and unlabeled images. Existing uncertainty-based semi-supervised methods [5,13,14] have achieved stunning results by considering the reliability of the supervision for the unlabeled images. These methods exploit the epistemic uncertainty, a kind of uncertainty about the model's parameters arising from a lack of data, either in the output space [5,13,14] or in the feature space [14], as guidance for identifying trustworthy supervision.…”

Section: Introductionmentioning

confidence: 99%

Exploring Feature Representation Learning for Semi-supervised Medical Image Segmentation

Wu¹,

Li²,

Cheng³

2021

Preprint

View full text Add to dashboard Cite

This paper presents a simple yet effective two-stage framework for semi-supervised medical image segmentation. Our key insight is to explore the feature representation learning with labeled and unlabeled (i.e., pseudo labeled) images to enhance the segmentation performance. In the first stage, we present an aleatoric uncertainty-aware method, namely AUA, to improve the segmentation performance for generating high-quality pseudo labels. Considering the inherent ambiguity of medical images, AUA adaptively regularizes the consistency on images with low ambiguity. To enhance the representation learning, we propose a stage-adaptive contrastive learning method, including a boundary-aware contrastive loss to regularize the labeled images in the first stage and a prototype-aware contrastive loss to optimize both labeled and pseudo labeled images in the second stage. The boundary-aware contrastive loss only optimizes pixels around the segmentation boundaries to reduce the computational cost. The prototype-aware contrastive loss fully leverages both labeled images and pseudo labeled images by building a centroid for each class to reduce computational cost for pair-wise comparison. Our method achieves the best results on two public medical image segmentation benchmarks. Notably, our method outperforms the prior state-of-the-art by 5.7% on Dice for colon tumor segmentation relying on just 5% labeled images.

show abstract

Double-Uncertainty Weighted Method for Semi-supervised Learning

Cited by 81 publications

References 7 publications

Two-Stage Segmentation Framework Based on Distance Transformation

Two-Stage Segmentation Framework Based on Distance Transformation

Uncertainty-Guided Mutual Consistency Learning for Semi-Supervised Medical Image Segmentation

Exploring Feature Representation Learning for Semi-supervised Medical Image Segmentation

Contact Info

Product

Resources

About