TransBridge: A Lightweight Transformer for Left Ventricle Segmentation in Echocardiography



Deng, Kaizhong, Meng, Yanda ORCID: 0000-0001-7344-2174, Gao, Dongxu ORCID: 0000-0001-7008-0737, Bridge, Joshua, Shen, Yaochun ORCID: 0000-0002-8915-1993, Lip, Gregory ORCID: 0000-0002-7566-1626, Zhao, Yitian and Zheng, Yalin ORCID: 0000-0002-7873-0922
(2021) TransBridge: A Lightweight Transformer for Left Ventricle Segmentation in Echocardiography. .

[img] Text
TransBridge (1).pdf - Author Accepted Manuscript

Download (460kB) | Preview

Abstract

Echocardiography is an essential diagnostic method to assess cardiac functions. However, manually labelling the left ventricle region on echocardiography images is time-consuming and subject to observer bias. Therefore, it is vital to develop a high-performance and efficient automatic assessment tool. Inspired by the success of the transformer structure in vision tasks, we develop a lightweight model named ‘TransBridge’ for segmentation tasks. This hybrid framework combines a convolutional neural network (CNN) encoder-decoder structure and a transformer structure. The transformer layers bridge the CNN encoder and decoder to fuse the multi-level features extracted by the CNN encoder, to build global and inter-level dependencies. A new patch embedding layer has been implemented using the dense patch division method and shuffled group convolution to reduce the excessive parameter number in the embedding layer and the size of the token sequence. The model is evaluated on the EchoNet-Dynamic dataset for the left ventricle segmentation task. The experimental results show that the total number of parameters is reduced by 78.7% compared to CoTr [22] and the Dice coefficient reaches 91.4%, proving the structure’s effectiveness.

Item Type: Conference or Workshop Item (Unspecified)
Uncontrolled Keywords: Echocardiography, Left ventricle segmentation, Lightweight transformer model, Parameter efficiency
Divisions: Faculty of Health and Life Sciences
Faculty of Health and Life Sciences > Institute of Life Courses and Medical Sciences
Depositing User: Symplectic Admin
Date Deposited: 20 Oct 2021 10:29
Last Modified: 22 Nov 2023 18:40
DOI: 10.1007/978-3-030-87583-1_7
Related URLs:
URI: https://livrepository.liverpool.ac.uk/id/eprint/3140960