Hybrid deep learning network for vascular segmentation in photoacoustic imaging

Alan Yilun Yuan; Yang Gao; Liangliang Peng; Lingxiao Zhou; Jun Liu; Siwei Zhu; Wei Song

doi:10.1364/BOE.409246

Hybrid deep learning network for vascular segmentation in photoacoustic imaging

Biomed Opt Express. 2020 Oct 16;11(11):6445-6457. doi: 10.1364/BOE.409246. eCollection 2020 Nov 1.

Authors

Alan Yilun Yuan^{1

2}, Yang Gao^{3

2}, Liangliang Peng³, Lingxiao Zhou^{4

5

6}, Jun Liu^{7

8}, Siwei Zhu⁷, Wei Song^{4

9}

Affiliations

¹ Department of Electrical and Electronic Engineering, Imperial College London, London, UK.
² These authors contributed equally to this work.
³ Nanophotonics Research Center, College of Physics and Optoelectronic Engineering, Shenzhen University, Shenzhen, China.
⁴ Nanophotonics Research Center, Shenzhen Key Laboratory of Micro-Scale Optical Information Technology, Institute of Microscale Optoelectronics, Shenzhen University, Shenzhen, China.
⁵ Department of Respiratory Medicine, Zhongshan-Xuhui Hospital, Fudan University, Shanghai, China.
⁶ lingxiaoz@szu.edu.cn.
⁷ Tianjin Union Medical Centre, Tianjin, China.
⁸ cjr.liujun@vip.163.com.
⁹ weisong@szu.edu.cn.

Abstract

Photoacoustic (PA) technology has been used extensively on vessel imaging due to its capability of identifying molecular specificities and achieving high optical-diffraction-limited lateral resolution down to the cellular level. Vessel images carry essential medical information that provides guidelines for a professional diagnosis. Modern image processing techniques provide a decent contribution to vessel segmentation. However, these methods suffer from under or over-segmentation. Thus, we demonstrate both the results of adopting a fully convolutional network and U-net, and propose a hybrid network consisting of both applied on PA vessel images. Comparison results indicate that the hybrid network can significantly increase the segmentation accuracy and robustness.