Elsevier

Photoacoustics

Volume 25, March 2022, 100310
Photoacoustics

Full-view in vivo skin and blood vessels profile segmentation in photoacoustic imaging based on deep learning

https://doi.org/10.1016/j.pacs.2021.100310Get rights and content
Under a Creative Commons license
open access

Abstract

Photoacoustic (PA) microscopy allows imaging of the soft biological tissue based on optical absorption contrast and spatial ultrasound resolution. One of the major applications of PA imaging is its characterization of microvasculature. However, the strong PA signal from skin layer overshadowed the subcutaneous blood vessels leading to indirectly reconstruct the PA images in human study. Addressing the present situation, we examined a deep learning (DL) automatic algorithm to achieve high-resolution and high-contrast segmentation for widening PA imaging applications. In this research, we propose a DL model based on modified U-Net for extracting the relationship features between amplitudes of the generated PA signal from skin and underlying vessels. This study illustrates the broader potential of hybrid complex network as an automatic segmentation tool for the in vivo PA imaging. With DL-infused solution, our result outperforms the previous studies with achieved real-time semantic segmentation on large-size high-resolution PA images.

Keywords

Photoacoustic imaging
Segmentation
High resolution
Deep learning
U-Net

Data availability

The python code in Keras with a TensorFlow backend implementation of the Slide-U-Net and all of the training datasets used for this study were acquired by our laboratory are available on Github: https://github.com/lycaoduong/ohlabs_pam_segmentation_unet.

Cited by (0)

Cao Duong Ly received the B.S. degree in Biomedical Engineering from Ho Chi Minh City University of Technology, Vietnam (2018) and M.S. degree from Pukyong National University, Republic of Korea (2020). He is currently working as a vision engineer and research assistant at Ohlabs Corporation, Busan, Republic of Korea. His research focuses on computer vision for biomedical application and industrial application.

Van Tu Nguyen received the M.Sc. degree in Biomedical Engineering from Pukyong National University, Republic of Korea in 2018. He is currently doing Ph.D. degree in Biomedical Engineering, Pukyong National University. His research focuses on biomedical imaging methods, including photoacoustic imaging and fluorescence imaging.

Tan Hung Vo received the B.S. degree in Ho Chi Minh University of Science, Vietnam (2014). He is currently pursuing the M.S. degree in Biomedical Engineering at Pukyong National University. His research interests Intelligent control based on neural network, Digital image processing.

Sudip Mondal obtained his Ph.D. in 2015 from CSIR-Central Mechanical Engineering Research Institute and National Institute Technology Durgapur, India. He joined as a Post-Doctoral Fellow at Benemérita Universidad Autónoma de Puebla (BUAP) University, Mexico (2015–2017). Currently he works as Research Professor, at the Department of Biomedical Engineering, in Pukyong National University, South Korea. His research interests include nanostructured materials synthesis, bioimaging, and biomedical applications such as cancer therapy and tissue engineering.

Sumin Park received the B.S. degree in Biomedical engineering, Pukyong National University, South Korea (2020). She is currently pursuing her M.S. degree in Industry 4.0 Convergence Bionics Engineering, Pukyong National University, South Korea. Her research interests include fabrication of the transducers, Scanning Acoustic Microscopy, hydroxyapatite and nanoparticles.

Jaeyeop Choi received the B.S. and M.S. degree in Biomedical Engineering from Pukyong National University, Busan, Republic of Korea. He is currently doing Ph.D. degree in Industry 4.0 Convergence Bionics Engineering, Pukyong National University. His current research interests include high frequency ultrasonic transducer and scanning acoustic microscopy.

Thi Thu Ha Vu is pursuing a master’s degree in Industry 4.0 Convergence Bionics Engineering, Pukyong National University, Republic of Korea. Her research focuses on applying Artificial Intelligence, Machine Learning, Deep Learning to Biomedical Image Processing. In addition, she is also working in Image processing in the industrial field.

Chang-Seok Kim received the Ph.D. degree from The Johns Hopkins University, Baltimore, MD, USA, in 2004. He is a Professor with the Department of Optics and Mechatronics Engineering and the Department of Cogno-Mechatronics Engineering, Pusan National University, Busan, South Korea. His current research interests include the development of novel fiber laser systems and application of them into biomedical, telecommunication, and sensor areas.

Junghwan Oh received the B.S. degree in Mechanical engineering from Pukyong National University in 1992, and the M.S. and Ph.D. degrees in Biomedical engineering from The University of Texas at Austin, USA, in 2003 and 2007, respectively. In 2010, he joined the Department of Biomedical Engineering at Pukyong National University, where he is a Full Professor. He also serves as Director of OhLabs Corporation research center. His current research interests include ultrasonic-based diagnostic imaging modalities for biomedical engineering applications, biomedical signal processing and health care systems.

1

These authors contributed equally to this work.