A Hybrid Robust-Learning Architecture for Medical Image Segmentation with Noisy Labels
Deep-learning models require large amounts of accurately labeled data. However, for medical image segmentation, high-quality labels rely on expert experience, and less-experienced operators provide noisy labels. How one might mitigate the negative effects caused by noisy labels for 3D medical image...
| Main Authors: | , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
MDPI AG
2022-01-01
|
| Series: | Future Internet |
| Subjects: | |
| Online Access: | https://www.mdpi.com/1999-5903/14/2/41 |
| _version_ | 1827654882655469568 |
|---|---|
| author | Jialin Shi Chenyi Guo Ji Wu |
| author_facet | Jialin Shi Chenyi Guo Ji Wu |
| author_sort | Jialin Shi |
| collection | DOAJ |
| description | Deep-learning models require large amounts of accurately labeled data. However, for medical image segmentation, high-quality labels rely on expert experience, and less-experienced operators provide noisy labels. How one might mitigate the negative effects caused by noisy labels for 3D medical image segmentation has not been fully investigated. In this paper, our purpose is to propose a novel hybrid robust-learning architecture to combat noisy labels for 3D medical image segmentation. Our method consists of three components. First, we focus on the noisy annotations of slices and propose a slice-level label-quality awareness method, which automatically generates label-quality scores for slices in a set. Second, we propose a shape-awareness regularization loss based on distance transform maps to introduce prior shape information and provide extra performance gains. Third, based on a re-weighting strategy, we propose an end-to-end hybrid robust-learning architecture to weaken the negative effects caused by noisy labels. Extensive experiments are performed on two representative datasets (i.e., liver segmentation and multi-organ segmentation). Our hybrid noise-robust architecture has shown competitive performance, compared to other methods. Ablation studies also demonstrate the effectiveness of slice-level label-quality awareness and a shape-awareness regularization loss for combating noisy labels. |
| first_indexed | 2024-03-09T21:55:05Z |
| format | Article |
| id | doaj.art-b7b8717f218642319f55ff28ca8c4ed0 |
| institution | Directory Open Access Journal |
| issn | 1999-5903 |
| language | English |
| last_indexed | 2024-03-09T21:55:05Z |
| publishDate | 2022-01-01 |
| publisher | MDPI AG |
| record_format | Article |
| series | Future Internet |
| spelling | doaj.art-b7b8717f218642319f55ff28ca8c4ed02023-11-23T19:59:46ZengMDPI AGFuture Internet1999-59032022-01-011424110.3390/fi14020041A Hybrid Robust-Learning Architecture for Medical Image Segmentation with Noisy LabelsJialin Shi0Chenyi Guo1Ji Wu2The Department of Electronic Engineering, Tsinghua University, Beijing 100089, ChinaThe Department of Electronic Engineering, Tsinghua University, Beijing 100089, ChinaThe Department of Electronic Engineering, Tsinghua University, Beijing 100089, ChinaDeep-learning models require large amounts of accurately labeled data. However, for medical image segmentation, high-quality labels rely on expert experience, and less-experienced operators provide noisy labels. How one might mitigate the negative effects caused by noisy labels for 3D medical image segmentation has not been fully investigated. In this paper, our purpose is to propose a novel hybrid robust-learning architecture to combat noisy labels for 3D medical image segmentation. Our method consists of three components. First, we focus on the noisy annotations of slices and propose a slice-level label-quality awareness method, which automatically generates label-quality scores for slices in a set. Second, we propose a shape-awareness regularization loss based on distance transform maps to introduce prior shape information and provide extra performance gains. Third, based on a re-weighting strategy, we propose an end-to-end hybrid robust-learning architecture to weaken the negative effects caused by noisy labels. Extensive experiments are performed on two representative datasets (i.e., liver segmentation and multi-organ segmentation). Our hybrid noise-robust architecture has shown competitive performance, compared to other methods. Ablation studies also demonstrate the effectiveness of slice-level label-quality awareness and a shape-awareness regularization loss for combating noisy labels.https://www.mdpi.com/1999-5903/14/2/41deep learningnoisy labelsmedical image segmentationrobust learning |
| spellingShingle | Jialin Shi Chenyi Guo Ji Wu A Hybrid Robust-Learning Architecture for Medical Image Segmentation with Noisy Labels Future Internet deep learning noisy labels medical image segmentation robust learning |
| title | A Hybrid Robust-Learning Architecture for Medical Image Segmentation with Noisy Labels |
| title_full | A Hybrid Robust-Learning Architecture for Medical Image Segmentation with Noisy Labels |
| title_fullStr | A Hybrid Robust-Learning Architecture for Medical Image Segmentation with Noisy Labels |
| title_full_unstemmed | A Hybrid Robust-Learning Architecture for Medical Image Segmentation with Noisy Labels |
| title_short | A Hybrid Robust-Learning Architecture for Medical Image Segmentation with Noisy Labels |
| title_sort | hybrid robust learning architecture for medical image segmentation with noisy labels |
| topic | deep learning noisy labels medical image segmentation robust learning |
| url | https://www.mdpi.com/1999-5903/14/2/41 |
| work_keys_str_mv | AT jialinshi ahybridrobustlearningarchitectureformedicalimagesegmentationwithnoisylabels AT chenyiguo ahybridrobustlearningarchitectureformedicalimagesegmentationwithnoisylabels AT jiwu ahybridrobustlearningarchitectureformedicalimagesegmentationwithnoisylabels AT jialinshi hybridrobustlearningarchitectureformedicalimagesegmentationwithnoisylabels AT chenyiguo hybridrobustlearningarchitectureformedicalimagesegmentationwithnoisylabels AT jiwu hybridrobustlearningarchitectureformedicalimagesegmentationwithnoisylabels |