Deep learning alignment of bidirectional raster scanning in high speed photoacoustic microscopy

Abstract

Simultaneous point-by-point raster scanning of optical and acoustic beams has been widely adapted to high-speed photoacoustic microscopy (PAM) using a water-immersible microelectromechanical system or galvanometer scanner. However, when using high-speed water-immersible scanners, the two consecutively acquired bidirectional PAM images are misaligned with each other because of unstable performance, which causes a non-uniform time interval between scanning points. Therefore, only one unidirectionally acquired image is typically used; consequently, the imaging speed is reduced by half. Here, we demonstrate a scanning framework based on a deep neural network (DNN) to correct misaligned PAM images acquired via bidirectional raster scanning. The proposed method doubles the imaging speed compared to that of conventional methods by aligning nonlinear mismatched cross-sectional B-scan photoacoustic images during bidirectional raster scanning. Our DNN-assisted raster scanning framework can further potentially be applied to other raster scanning-based biomedical imaging tools, such as optical coherence tomography, ultrasound microscopy, and confocal microscopy.

High-speed photoacoustic monitoring of vascular changes during acute hyperglycemia

“…The use of the CGM system could replace the invasive BG meter, which would facilitate these kinds of experiments. Additionally, system improvement for high resolution and high speed, integration with other sensing and imaging modalities, and advanced signal and image processing including nonlinear filters and deep learning could be applied to obtain better results [19][20][21][22][23][24][25]. The experimental results introduced in this paper are insufficient in the number of samples and depth of analysis.…”

Recent Advances in Photoacoustic Agents for Theranostic Applications

“…Recently, three-dimensional PA computed tomography (PACT) systems have shown great potential for monitoring the biodistribution of small animals in vivo [ 26 ]. In addition, advanced deep-learning algorithms have been reported to achieve better image quality [ 99 , 100 , 101 , 102 , 103 , 104 ]. The continuous improvement of the PAI system can potentially translate the PAI-guided phototherapy technique to clinical applications.…”

Sounding out the dynamics: a concise review of high-speed photoacoustic microscopy

“… Improving PAM images by (a) aligning polygon facets in a fast polygon scanning system combined with deep learning FD U-Net upsampling, 22 (b) recovering misaligned signal from bidirectional scan using deep learning, 37 (c) correcting MEMS scanning distortion using a deep learning spatial weight matrix (SWM) with a dimensionality reduction, 38 and (d) improving the image quality of low energy PAM using MT-RDN. 39 Images (a) and (b)–(d) are reprinted with permission from Refs.…”

“…Other computational techniques speed up PAM image formation by attempting to correct imperfections in the scanning mechanism, such as by quickly restoring misaligned/distorted aspects of the scanning path or denoising low-energy images acquired within certain laser dosage limits. 22 , 37 – 39 …”