Masked Modeling-Based Ultrasound Image Classification via Self-Supervised Learning
Recently, deep learning-based methods have emerged as the preferred approach for ultrasound data analysis. However, these methods often require large-scale annotated datasets for training deep models, which are not readily available in practical scenarios. Additionally, the presence of speckle noise...
Saved in:
| Main Authors: | , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
IEEE
2024-01-01
|
| Series: | IEEE Open Journal of Engineering in Medicine and Biology |
| Subjects: | |
| Online Access: | https://ieeexplore.ieee.org/document/10463101/ |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1832582308355375104 |
|---|---|
| author | Kele Xu Kang You Boqing Zhu Ming Feng Dawei Feng Cheng Yang |
| author_facet | Kele Xu Kang You Boqing Zhu Ming Feng Dawei Feng Cheng Yang |
| author_sort | Kele Xu |
| collection | DOAJ |
| description | Recently, deep learning-based methods have emerged as the preferred approach for ultrasound data analysis. However, these methods often require large-scale annotated datasets for training deep models, which are not readily available in practical scenarios. Additionally, the presence of speckle noise and other imaging artifacts can introduce numerous hard examples for ultrasound data classification. In this paper, drawing inspiration from self-supervised learning techniques, we present a pre-training method based on mask modeling specifically designed for ultrasound data. Our study investigates three different mask modeling strategies: random masking, vertical masking, and horizontal masking. By employing these strategies, our pre-training approach aims to predict the masked portion of the ultrasound images. Notably, our method does not rely on externally labeled data, allowing us to extract representative features without the need for human annotation. Consequently, we can leverage unlabeled datasets for pre-training. Furthermore, to address the challenges posed by hard samples in ultrasound data, we propose a novel hard sample mining strategy. To evaluate the effectiveness of our proposed method, we conduct experiments on two datasets. The experimental results demonstrate that our approach outperforms other state-of-the-art methods in ultrasound image classification. This indicates the superiority of our pre-training method and its ability to extract discriminative features from ultrasound data, even in the presence of hard examples. |
| format | Article |
| id | doaj-art-6f9220eb1a9748cab67b3e7e517e7665 |
| institution | Kabale University |
| issn | 2644-1276 |
| language | English |
| publishDate | 2024-01-01 |
| publisher | IEEE |
| record_format | Article |
| series | IEEE Open Journal of Engineering in Medicine and Biology |
| spelling | doaj-art-6f9220eb1a9748cab67b3e7e517e76652025-01-30T00:03:32ZengIEEEIEEE Open Journal of Engineering in Medicine and Biology2644-12762024-01-01522623710.1109/OJEMB.2024.337496610463101Masked Modeling-Based Ultrasound Image Classification via Self-Supervised LearningKele Xu0https://orcid.org/0000-0001-5997-5169Kang You1https://orcid.org/0009-0009-5685-2521Boqing Zhu2https://orcid.org/0000-0001-7867-2112Ming Feng3https://orcid.org/0000-0001-8062-6246Dawei Feng4https://orcid.org/0000-0002-7587-8905Cheng Yang5https://orcid.org/0000-0002-4782-1645National University of Defense Technology, Changsha, ChinaShanghai Jiao Tong University, Shanghai, ChinaNational University of Defense Technology, Changsha, ChinaTongJi University, Shanghai, ChinaNational University of Defense Technology, Changsha, ChinaNational University of Defense Technology, Changsha, ChinaRecently, deep learning-based methods have emerged as the preferred approach for ultrasound data analysis. However, these methods often require large-scale annotated datasets for training deep models, which are not readily available in practical scenarios. Additionally, the presence of speckle noise and other imaging artifacts can introduce numerous hard examples for ultrasound data classification. In this paper, drawing inspiration from self-supervised learning techniques, we present a pre-training method based on mask modeling specifically designed for ultrasound data. Our study investigates three different mask modeling strategies: random masking, vertical masking, and horizontal masking. By employing these strategies, our pre-training approach aims to predict the masked portion of the ultrasound images. Notably, our method does not rely on externally labeled data, allowing us to extract representative features without the need for human annotation. Consequently, we can leverage unlabeled datasets for pre-training. Furthermore, to address the challenges posed by hard samples in ultrasound data, we propose a novel hard sample mining strategy. To evaluate the effectiveness of our proposed method, we conduct experiments on two datasets. The experimental results demonstrate that our approach outperforms other state-of-the-art methods in ultrasound image classification. This indicates the superiority of our pre-training method and its ability to extract discriminative features from ultrasound data, even in the presence of hard examples.https://ieeexplore.ieee.org/document/10463101/Pre-trainingself-supervisedultrasound imagemasked modeling |
| spellingShingle | Kele Xu Kang You Boqing Zhu Ming Feng Dawei Feng Cheng Yang Masked Modeling-Based Ultrasound Image Classification via Self-Supervised Learning IEEE Open Journal of Engineering in Medicine and Biology Pre-training self-supervised ultrasound image masked modeling |
| title | Masked Modeling-Based Ultrasound Image Classification via Self-Supervised Learning |
| title_full | Masked Modeling-Based Ultrasound Image Classification via Self-Supervised Learning |
| title_fullStr | Masked Modeling-Based Ultrasound Image Classification via Self-Supervised Learning |
| title_full_unstemmed | Masked Modeling-Based Ultrasound Image Classification via Self-Supervised Learning |
| title_short | Masked Modeling-Based Ultrasound Image Classification via Self-Supervised Learning |
| title_sort | masked modeling based ultrasound image classification via self supervised learning |
| topic | Pre-training self-supervised ultrasound image masked modeling |
| url | https://ieeexplore.ieee.org/document/10463101/ |
| work_keys_str_mv | AT kelexu maskedmodelingbasedultrasoundimageclassificationviaselfsupervisedlearning AT kangyou maskedmodelingbasedultrasoundimageclassificationviaselfsupervisedlearning AT boqingzhu maskedmodelingbasedultrasoundimageclassificationviaselfsupervisedlearning AT mingfeng maskedmodelingbasedultrasoundimageclassificationviaselfsupervisedlearning AT daweifeng maskedmodelingbasedultrasoundimageclassificationviaselfsupervisedlearning AT chengyang maskedmodelingbasedultrasoundimageclassificationviaselfsupervisedlearning |