High-fidelity deconvolution for acoustic-resolution photoacoustic microscopy enabled by convolutional neural networks

Fei Feng; Siqi Liang; Jiajia Luo; Sung-Liang Chen

doi:10.1016/j.pacs.2022.100360

High-fidelity deconvolution for acoustic-resolution photoacoustic microscopy enabled by convolutional neural networks

Photoacoustics. 2022 Apr 26:26:100360. doi: 10.1016/j.pacs.2022.100360. eCollection 2022 Jun.

Authors

Fei Feng¹, Siqi Liang¹, Jiajia Luo^{2

3

4}, Sung-Liang Chen^{1

5

6}

Affiliations

¹ University of Michigan-Shanghai Jiao Tong University Joint Institute, Shanghai Jiao Tong University, Shanghai 200240, China.
² Institute of Medical Technology, Peking University Health Science Center, Beijing 100191, China.
³ Biomedical Engineering Department, Peking University, Beijing 100191, China.
⁴ Peking University People's Hospital, Beijing 100044, China.
⁵ State Key Laboratory of Advanced Optical Communication Systems and Networks, Shanghai Jiao Tong University, Shanghai 200240, China.
⁶ Engineering Research Center of Digital Medicine and Clinical Translation, Ministry of Education, Shanghai 200030, China.

Abstract

Acoustic-resolution photoacoustic microscopy (AR-PAM) image resolution is determined by the point spread function (PSF) of the imaging system. Previous algorithms, including Richardson-Lucy (R-L) deconvolution and model-based (MB) deconvolution, improve spatial resolution by taking advantage of the PSF as prior knowledge. However, these methods encounter the problems of inaccurate deconvolution, meaning the deconvolved feature size and the original one are not consistent (e.g., the former can be smaller than the latter). We present a novel deep convolution neural network (CNN)-based algorithm featuring high-fidelity recovery of multiscale feature size to improve lateral resolution of AR-PAM. The CNN is trained with simulated image pairs of line patterns, which is to mimic blood vessels. To investigate the suitable CNN model structure and elaborate on the effectiveness of CNN methods compared with non-learning methods, we select five different CNN models, while R-L and directional MB methods are also applied for comparison. Besides simulated data, experimental data including tungsten wires, leaf veins, and in vivo blood vessels are also evaluated. A custom-defined metric of relative size error (RSE) is used to quantify the multiscale feature recovery ability of different methods. Compared to other methods, enhanced deep super resolution (EDSR) network and residual in residual dense block network (RRDBNet) model show better recovery in terms of RSE for tungsten wires with diameters ranging from 30 $μ m$ to 120 $μ m$ . Moreover, AR-PAM images of leaf veins are tested to demonstrate the effectiveness of the optimized CNN methods (by EDSR and RRDBNet) for complex patterns. Finally, in vivo images of mouse ear blood vessels and rat ear blood vessels are acquired and then deconvolved, and the results show that the proposed CNN method (notably RRDBNet) enables accurate deconvolution of multiscale feature size and thus good fidelity.

Keywords: Deconvolution; Deep learning; High-fidelity deconvolution; Multiscale imaging; Photoacoustic imaging.