Unsupervised Denoising of Photoacoustic Images Based on the Noise2Noise Network

Overview

Journal Biomed Opt Express

Specialty Radiology

Date 2024 Sep 30

PMID 39346987

Authors

Yanda Cheng

Wenhan Zheng

Robert Bing

Huijuan Zhang

Chuqin Huang

Peizhou Huang

Leslie Ying

Jun Xia

Affiliations

Soon will be listed here.

Abstract

In this study, we implemented an unsupervised deep learning method, the Noise2Noise network, for the improvement of linear-array-based photoacoustic (PA) imaging. Unlike supervised learning, which requires a noise-free ground truth, the Noise2Noise network can learn noise patterns from a pair of noisy images. This is particularly important for in vivo PA imaging, where the ground truth is not available. In this study, we developed a method to generate noise pairs from a single set of PA images and verified our approach through simulation and experimental studies. Our results reveal that the method can effectively remove noise, improve signal-to-noise ratio, and enhance vascular structures at deeper depths. The denoised images show clear and detailed vascular structure at different depths, providing valuable insights for preclinical research and potential clinical applications.

Citing Articles

Enhanced clinical photoacoustic vascular imaging through a skin localization network and adaptive weighting.

Huang C, Zheng E, Zheng W, Zhang H, Cheng Y, Zhang X Photoacoustics. 2025; 42:100690.

PMID: 39916976 PMC: 11800082. DOI: 10.1016/j.pacs.2025.100690.

References

Holan S, Viator J . Automated wavelet denoising of photoacoustic signals for circulating melanoma cell detection and burn image reconstruction. Phys Med Biol. 2008; 53(12):N227-36. DOI: 10.1088/0031-9155/53/12/N01. View

Zhu N, Liu C, Forsyth B, Singer Z, Laine A, Danino T . Segmentation with Residual Attention U-Net and an Edge-Enhancement Approach Preserves Cell Shape Features. Annu Int Conf IEEE Eng Med Biol Soc. 2022; 2022:2115-2118. DOI: 10.1109/EMBC48229.2022.9871026. View

Mallidi S, Emelianov S . Photoacoustic technique to measure beam profile of pulsed laser systems. Rev Sci Instrum. 2009; 80(5):054901. PMC: 2736568. DOI: 10.1063/1.3125625. View

Wang Y, Zhan Y, Harris L, Khan S, Xia J . A portable three-dimensional photoacoustic tomography system for imaging of chronic foot ulcers. Quant Imaging Med Surg. 2019; 9(5):799-806. PMC: 6571196. DOI: 10.21037/qims.2019.05.02. View

Deng H, Qiao H, Dai Q, Ma C . Deep learning in photoacoustic imaging: a review. J Biomed Opt. 2021; 26(4). PMC: 8033250. DOI: 10.1117/1.JBO.26.4.040901. View

Feng Z, Nie D, Wang L, Shen D . SEMI-SUPERVISED LEARNING FOR PELVIC MR IMAGE SEGMENTATION BASED ON MULTI-TASK RESIDUAL FULLY CONVOLUTIONAL NETWORKS. Proc IEEE Int Symp Biomed Imaging. 2018; 2018:885-888. PMC: 6193482. DOI: 10.1109/ISBI.2018.8363713. View

Dehner C, Olefir I, Chowdhury K, Justel D, Ntziachristos V . Deep-Learning-Based Electrical Noise Removal Enables High Spectral Optoacoustic Contrast in Deep Tissue. IEEE Trans Med Imaging. 2022; 41(11):3182-3193. DOI: 10.1109/TMI.2022.3180115. View

Bian W, Jang A, Liu F . Improving quantitative MRI using self-supervised deep learning with model reinforcement: Demonstration for rapid T1 mapping. Magn Reson Med. 2024; 92(1):98-111. PMC: 11055673. DOI: 10.1002/mrm.30045. View

Zackrisson S, van de Ven S, Gambhir S . Light in and sound out: emerging translational strategies for photoacoustic imaging. Cancer Res. 2014; 74(4):979-1004. PMC: 3944207. DOI: 10.1158/0008-5472.CAN-13-2387. View

10.

Gulenko O, Yang H, Kim K, Youm J, Kim M, Kim Y . Deep-Learning-Based Algorithm for the Removal of Electromagnetic Interference Noise in Photoacoustic Endoscopic Image Processing. Sensors (Basel). 2022; 22(10). PMC: 9147354. DOI: 10.3390/s22103961. View

11.

Kazakeviciute A, Ho C, Olivo M . Multispectral Photoacoustic Imaging Artifact Removal and Denoising Using Time Series Model-Based Spectral Noise Estimation. IEEE Trans Med Imaging. 2016; 35(9):2151-2163. DOI: 10.1109/TMI.2016.2550624. View

12.

Zheng W, Zhang H, Huang C, Shijo V, Xu C, Xu W . Deep Learning Enhanced Volumetric Photoacoustic Imaging of Vasculature in Human. Adv Sci (Weinh). 2023; 10(29):e2301277. PMC: 10582405. DOI: 10.1002/advs.202301277. View

13.

Martinez-Perez M, Hughes A, Thom S, Parker K . Improvement of a retinal blood vessel segmentation method using the Insight Segmentation and Registration Toolkit (ITK). Annu Int Conf IEEE Eng Med Biol Soc. 2007; 2007:892-5. DOI: 10.1109/IEMBS.2007.4352434. View

14.

Guney G, Uluc N, Demirkiran A, Aytac-Kipergil E, Unlu M, Birgul O . Comparison of noise reduction methods in photoacoustic microscopy. Comput Biol Med. 2019; 109:333-341. DOI: 10.1016/j.compbiomed.2019.04.035. View

15.

Bernhardt M, Castro D, Tanno R, Schwaighofer A, Tezcan K, Monteiro M . Active label cleaning for improved dataset quality under resource constraints. Nat Commun. 2022; 13(1):1161. PMC: 8897392. DOI: 10.1038/s41467-022-28818-3. View

16.

Zhang L, Zhang L, Mou X, Zhang D . FSIM: a feature similarity index for image quality assessment. IEEE Trans Image Process. 2011; 20(8):2378-86. DOI: 10.1109/TIP.2011.2109730. View

17.

Xia J, Yao J, Wang L . Photoacoustic tomography: principles and advances. Electromagn Waves (Camb). 2015; 147:1-22. PMC: 4311576. DOI: 10.2528/pier14032303. View

18.

Wang Y, Lim R, Zhang H, Nyayapathi N, Oh K, Xia J . Optimizing the light delivery of linear-array-based photoacoustic systems by double acoustic reflectors. Sci Rep. 2018; 8(1):13004. PMC: 6115359. DOI: 10.1038/s41598-018-31430-5. View

19.

Kaur P, Singh G, Kaur P . A Review of Denoising Medical Images Using Machine Learning  Approaches. Curr Med Imaging Rev. 2018; 14(5):675-685. PMC: 6225344. DOI: 10.2174/1573405613666170428154156. View

20.

Jung W, Lee H, Seo M, Nam Y, Choi Y, Shin N . MR-self Noise2Noise: self-supervised deep learning-based image quality improvement of submillimeter resolution 3D MR images. Eur Radiol. 2022; 33(4):2686-2698. DOI: 10.1007/s00330-022-09243-y. View