| 龚瑾儒,翟锦涛,田峰,雷卫瑞,常帅,王润民,邹孝,钱盛友.融合注意力和状态空间的儿科超声心动图分割[J].电子测量与仪器学报,2025,39(10):32-40 |
| 融合注意力和状态空间的儿科超声心动图分割 |
| Pediatric echocardiography segmentation combining attention and state space |
| |
| DOI： |
| 中文关键词: 深度学习 图像分割 儿科超声心动图 注意力机制 状态空间模型 |
| 英文关键词:deep learning image segmentation pediatric echocardiography attention mechanism state space model |
| 基金项目:国家自然科学基金（12274200，61502164）、湖南省教育厅科研基金（21A0052，22B0036）、湖南省自然科学基金（2020JJ4057）、长沙市自然科学基金（kq2202239）项目资助 |
|
|
| Author | Institution |
| Gong Jinru | School of Physics and Electronics, Hunan Normal University, Changsha 410081, China |
| Zhai Jintao | School of Physics and Electronics, Hunan Normal University, Changsha 410081, China |
| Tian Feng | School of Physics and Electronics, Hunan Normal University, Changsha 410081, China |
| Lei Weirui | School of Physics and Electronics, Hunan Normal University, Changsha 410081, China |
| Chang Shuai | School of Physics and Electronics, Hunan Normal University, Changsha 410081, China |
| Wang Runmin | School of Information Science and Engineering, Hunan Normal University, Changsha 410081, China |
| Zou Xiao | School of Physics and Electronics, Hunan Normal University, Changsha 410081, China |
| Qian Shengyou | School of Physics and Electronics, Hunan Normal University, Changsha 410081, China |
|
| 摘要点击次数: 576 |
| 全文下载次数: 490 |
| 中文摘要: |
| 不同年龄段的儿童心脏尺寸差异大且儿童心率过快会导致心脏边界相较成人更加模糊,影响超声心动图的分割效果。针对上述问题,对H2Former分割模型进行改进,提出了一种时间和位置注意力的分层混合视觉状态空间模型（TPA-H2VSS）对儿科超声心动图左心室进行分割。首先替换Transformer模块为视觉状态空间模块,改进医学图像分割的长期依赖性建模关系;其次在模型的编码器与解码器之间搭建时间注意力模块,把超声心动视频左心室的语义信息在时间维度上进行补充和交互;最后,在输出部分加入位置注意力模块,进一步提高网络的分割性能。在儿科超声心动视频数据集EchoNet-Pediatric的PSAX数据集和A4C数据集上分别进行训练、验证和测试。与基线模型H2Former相比,在PSAX数据集上的Dice、交并比（IoU）、准确率（accuracy）分别提升了0.86%、1.41%、0.15%,豪斯多夫距离（HD）降低了0.219 5;在A4C数据集上的Dice、IoU、Accuracy分别提升了0.93%、1.53%、0.2%,HD降低了0.167。与其他模型进行比较,该模型能有效分割儿科超声心动图左心室,可以为先天性心脏病辅助诊断提供新的解决方案。 |
| 英文摘要: |
| The significant variation in cardiac dimensions across different age groups and the faster heart rate in children result in more blurred cardiac borders compared to adults, impacting the segmentation of echocardiography. To address the above problems, the segmentation model called H2Former is improved, and the model called TPA-H2VSS combining the attention and state space is proposed to segment the left ventricle of pediatric echocardiography. Firstly, this paper replaces the Transformer block with the visual state space (VSS) block to enhance the model’s advantage in long-range modeling. Secondly, the temporal attention (TA) module is built between the encoder and decoder in the model to complements and interacts with the semantic information of the left ventricle in the echocardiography video in the temporal dimension. Finally, the positional attention (PA) module is added in the output head to make pediatric echocardiographic left ventricle segmentation more accurate. The experiments were trained, validated, and tested on the pediatric echocardiographic video dataset EchoNet-Pediatrics on the PSAX dataset and the A4C dataset, respectively. Compared with the base model H2Former, Dice, IoU, and accuracy on the PSAX dataset were improved by 0.86%, 1.41%, and 0.15%, respectively, and HD was reduced by 0.219 5. Dice, IoU, and accuracy on the A4C dataset were improved by 0.93%, 1.53%, and 0.2%, respectively, and HD was reduced by 0.167. By comparing with other models, it was demonstrated that the model could effectively segment the left ventricle in pediatric echocardiography and could provide a new solution for the auxiliary diagnosis of congenital heart disease. |
| 查看全文 查看/发表评论 下载PDF阅读器 |
|
|
|
