Left Ventricle Segmentation in Echocardiography with Transformer
Left ventricular ejection fraction (LVEF) plays as an essential role in the assessment of cardiac function, providing quantitative data support for the medical diagnosis of heart disease. Robust evaluation of the ejection fraction relies on accurate left ventricular (LV) segmentation of echocardiogr...
| Main Authors: | , , , , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
MDPI AG
2023-07-01
|
| Series: | Diagnostics |
| Subjects: | |
| Online Access: | https://www.mdpi.com/2075-4418/13/14/2365 |
| _version_ | 1827733346508079104 |
|---|---|
| author | Minqi Liao Yifan Lian Yongzhao Yao Lihua Chen Fei Gao Long Xu Xin Huang Xinxing Feng Suxia Guo |
| author_facet | Minqi Liao Yifan Lian Yongzhao Yao Lihua Chen Fei Gao Long Xu Xin Huang Xinxing Feng Suxia Guo |
| author_sort | Minqi Liao |
| collection | DOAJ |
| description | Left ventricular ejection fraction (LVEF) plays as an essential role in the assessment of cardiac function, providing quantitative data support for the medical diagnosis of heart disease. Robust evaluation of the ejection fraction relies on accurate left ventricular (LV) segmentation of echocardiograms. Because human bias and expensive labor cost exist in manual echocardiographic analysis, computer algorithms of deep-learning have been developed to help human experts in segmentation tasks. Most of the previous work is based on the convolutional neural networks (CNN) structure and has achieved good results. However, the region occupied by the left ventricle is large for echocardiography. Therefore, the limited receptive field of CNN leaves much room for improvement in the effectiveness of LV segmentation. In recent years, Vision Transformer models have demonstrated their effectiveness and universality in traditional semantic segmentation tasks. Inspired by this, we propose two models that use two different pure Transformers as the basic framework for LV segmentation in echocardiography: one combines Swin Transformer and K-Net, and the other uses Segformer. We evaluate these two models on the EchoNet-Dynamic dataset of LV segmentation and compare the quantitative metrics with other models for LV segmentation. The experimental results show that the mean Dice similarity of the two models scores are 92.92% and 92.79%, respectively, which outperform most of the previous mainstream CNN models. In addition, we found that for some samples that were not easily segmented, whereas both our models successfully recognized the valve region and separated left ventricle and left atrium, the CNN model segmented them together as a single part. Therefore, it becomes possible for us to obtain accurate segmentation results through simple post-processing, by filtering out the parts with the largest circumference or pixel square. These promising results prove the effectiveness of the two models and reveal the potential of Transformer structure in echocardiographic segmentation. |
| first_indexed | 2024-03-11T01:10:10Z |
| format | Article |
| id | doaj.art-d43832a7e4a142e4b3156e43881a379f |
| institution | Directory Open Access Journal |
| issn | 2075-4418 |
| language | English |
| last_indexed | 2024-03-11T01:10:10Z |
| publishDate | 2023-07-01 |
| publisher | MDPI AG |
| record_format | Article |
| series | Diagnostics |
| spelling | doaj.art-d43832a7e4a142e4b3156e43881a379f2023-11-18T18:57:45ZengMDPI AGDiagnostics2075-44182023-07-011314236510.3390/diagnostics13142365Left Ventricle Segmentation in Echocardiography with TransformerMinqi Liao0Yifan Lian1Yongzhao Yao2Lihua Chen3Fei Gao4Long Xu5Xin Huang6Xinxing Feng7Suxia Guo8Department of Cardiology, Dongguan People’s Hospital (The Tenth Affiliated Hospital of Southern Medical Univerity), No 78, Wandao Road, Wanjiang District, Dongguan 523059, ChinaNational Space Science Center, Chinese Academy of Sciences, Beijing 100190, ChinaDepartment of Cardiology, Dongguan People’s Hospital (The Tenth Affiliated Hospital of Southern Medical Univerity), No 78, Wandao Road, Wanjiang District, Dongguan 523059, ChinaDepartment of Cardiology, Dongguan People’s Hospital (The Tenth Affiliated Hospital of Southern Medical Univerity), No 78, Wandao Road, Wanjiang District, Dongguan 523059, ChinaNational Space Science Center, Chinese Academy of Sciences, Beijing 100190, ChinaNational Astronomical Observatories, Chinese Academy of Sciences, Beijing 100101, ChinaNational Space Science Center, Chinese Academy of Sciences, Beijing 100190, ChinaEndocrinology Centre, Fuwai Hospital, Chinese Academy of Medical Sciences and Peking Union Medical College, Beijing 100037, ChinaDepartment of Cardiology, Dongguan People’s Hospital (The Tenth Affiliated Hospital of Southern Medical Univerity), No 78, Wandao Road, Wanjiang District, Dongguan 523059, ChinaLeft ventricular ejection fraction (LVEF) plays as an essential role in the assessment of cardiac function, providing quantitative data support for the medical diagnosis of heart disease. Robust evaluation of the ejection fraction relies on accurate left ventricular (LV) segmentation of echocardiograms. Because human bias and expensive labor cost exist in manual echocardiographic analysis, computer algorithms of deep-learning have been developed to help human experts in segmentation tasks. Most of the previous work is based on the convolutional neural networks (CNN) structure and has achieved good results. However, the region occupied by the left ventricle is large for echocardiography. Therefore, the limited receptive field of CNN leaves much room for improvement in the effectiveness of LV segmentation. In recent years, Vision Transformer models have demonstrated their effectiveness and universality in traditional semantic segmentation tasks. Inspired by this, we propose two models that use two different pure Transformers as the basic framework for LV segmentation in echocardiography: one combines Swin Transformer and K-Net, and the other uses Segformer. We evaluate these two models on the EchoNet-Dynamic dataset of LV segmentation and compare the quantitative metrics with other models for LV segmentation. The experimental results show that the mean Dice similarity of the two models scores are 92.92% and 92.79%, respectively, which outperform most of the previous mainstream CNN models. In addition, we found that for some samples that were not easily segmented, whereas both our models successfully recognized the valve region and separated left ventricle and left atrium, the CNN model segmented them together as a single part. Therefore, it becomes possible for us to obtain accurate segmentation results through simple post-processing, by filtering out the parts with the largest circumference or pixel square. These promising results prove the effectiveness of the two models and reveal the potential of Transformer structure in echocardiographic segmentation.https://www.mdpi.com/2075-4418/13/14/2365echocardiographyleft ventriclesegmentationtransformer |
| spellingShingle | Minqi Liao Yifan Lian Yongzhao Yao Lihua Chen Fei Gao Long Xu Xin Huang Xinxing Feng Suxia Guo Left Ventricle Segmentation in Echocardiography with Transformer Diagnostics echocardiography left ventricle segmentation transformer |
| title | Left Ventricle Segmentation in Echocardiography with Transformer |
| title_full | Left Ventricle Segmentation in Echocardiography with Transformer |
| title_fullStr | Left Ventricle Segmentation in Echocardiography with Transformer |
| title_full_unstemmed | Left Ventricle Segmentation in Echocardiography with Transformer |
| title_short | Left Ventricle Segmentation in Echocardiography with Transformer |
| title_sort | left ventricle segmentation in echocardiography with transformer |
| topic | echocardiography left ventricle segmentation transformer |
| url | https://www.mdpi.com/2075-4418/13/14/2365 |
| work_keys_str_mv | AT minqiliao leftventriclesegmentationinechocardiographywithtransformer AT yifanlian leftventriclesegmentationinechocardiographywithtransformer AT yongzhaoyao leftventriclesegmentationinechocardiographywithtransformer AT lihuachen leftventriclesegmentationinechocardiographywithtransformer AT feigao leftventriclesegmentationinechocardiographywithtransformer AT longxu leftventriclesegmentationinechocardiographywithtransformer AT xinhuang leftventriclesegmentationinechocardiographywithtransformer AT xinxingfeng leftventriclesegmentationinechocardiographywithtransformer AT suxiaguo leftventriclesegmentationinechocardiographywithtransformer |