Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

SingBAG Pro: Accelerating point cloud-based iterative reconstruction for 3D photoacoustic imaging under arbitrary array

Shuang Li, Yibing Wang, Jian Gao, Chulhong Kim, Seongwook Choi, Yu Zhang, Qian Chen, Yao Yao, Changhui Li

TL;DR

This work tackles 3D photoacoustic imaging with irregular, sparse transducer arrays by introducing SlingBAG Pro, a point-cloud iterative reconstruction framework that generalizes to arbitrary array geometries. It integrates zero-gradient filtering to initialize a compact point cloud and a hierarchical optimization strategy with progressively increasing temporal sampling rates to accelerate convergence. The approach yields up to a 2.2× speedup over the original SlingBAG while maintaining high image quality, demonstrated in simulations on hand-vessel phantoms and in vivo rat liver and kidney studies; the method also delivers superior CNR and artifact suppression compared with universal back-projection. The released code enables adoption for wearable and conformal array configurations, potentially enabling high-quality, compact 3D PAI in clinical and preclinical settings.

Abstract

High-quality three-dimensional (3D) photoacoustic imaging (PAI) is gaining increasing attention in clinical applications. To address the challenges of limited space and high costs, irregular geometric transducer arrays that conform to specific imaging regions are promising for achieving high-quality 3D PAI with fewer transducers. However, traditional iterative reconstruction algorithms struggle with irregular array configurations, suffering from high computational complexity, substantial memory requirements, and lengthy reconstruction times. In this work, we introduce SlingBAG Pro, an advanced reconstruction algorithm based on the point cloud iteration concept of the Sliding ball adaptive growth (SlingBAG) method, while extending its compatibility to arbitrary array geometries. SlingBAG Pro maintains high reconstruction quality, reduces the number of required transducers, and employs a hierarchical optimization strategy that combines zero-gradient filtering with progressively increased temporal sampling rates during iteration. This strategy rapidly removes redundant spatial point clouds, accelerates convergence, and significantly shortens overall reconstruction time. Compared to the original SlingBAG algorithm, SlingBAG Pro achieves up to a 2.2-fold speed improvement in point cloud-based 3D PA reconstruction under irregular array geometries. The proposed method is validated through both simulation and in vivo mouse experiments, and the source code is publicly available at https://github.com/JaegerCQ/SlingBAG_Pro.

SingBAG Pro: Accelerating point cloud-based iterative reconstruction for 3D photoacoustic imaging under arbitrary array

TL;DR

This work tackles 3D photoacoustic imaging with irregular, sparse transducer arrays by introducing SlingBAG Pro, a point-cloud iterative reconstruction framework that generalizes to arbitrary array geometries. It integrates zero-gradient filtering to initialize a compact point cloud and a hierarchical optimization strategy with progressively increasing temporal sampling rates to accelerate convergence. The approach yields up to a 2.2× speedup over the original SlingBAG while maintaining high image quality, demonstrated in simulations on hand-vessel phantoms and in vivo rat liver and kidney studies; the method also delivers superior CNR and artifact suppression compared with universal back-projection. The released code enables adoption for wearable and conformal array configurations, potentially enabling high-quality, compact 3D PAI in clinical and preclinical settings.

Abstract

High-quality three-dimensional (3D) photoacoustic imaging (PAI) is gaining increasing attention in clinical applications. To address the challenges of limited space and high costs, irregular geometric transducer arrays that conform to specific imaging regions are promising for achieving high-quality 3D PAI with fewer transducers. However, traditional iterative reconstruction algorithms struggle with irregular array configurations, suffering from high computational complexity, substantial memory requirements, and lengthy reconstruction times. In this work, we introduce SlingBAG Pro, an advanced reconstruction algorithm based on the point cloud iteration concept of the Sliding ball adaptive growth (SlingBAG) method, while extending its compatibility to arbitrary array geometries. SlingBAG Pro maintains high reconstruction quality, reduces the number of required transducers, and employs a hierarchical optimization strategy that combines zero-gradient filtering with progressively increased temporal sampling rates during iteration. This strategy rapidly removes redundant spatial point clouds, accelerates convergence, and significantly shortens overall reconstruction time. Compared to the original SlingBAG algorithm, SlingBAG Pro achieves up to a 2.2-fold speed improvement in point cloud-based 3D PA reconstruction under irregular array geometries. The proposed method is validated through both simulation and in vivo mouse experiments, and the source code is publicly available at https://github.com/JaegerCQ/SlingBAG_Pro.
Paper Structure (9 sections, 13 equations, 7 figures, 2 tables, 1 algorithm)

This paper contains 9 sections, 13 equations, 7 figures, 2 tables, 1 algorithm.

Table of Contents

  1. Introduction
  2. Results
  3. 3D PA image reconstruction under sparse irregular array
  4. 3D PA image reconstruction of in vivo animal studies with hierarchical optimization
  5. Discussion
  6. Methods
  7. Point cloud-based iterative reconstruction algorithm
  8. Zero-gradient filtering for point cloud initialization
  9. Hierarchical optimization based on variable sampling rates

Figures (7)

  • Figure 1: Comparison of 3D photoacoustic reconstruction results under sparse irregular arrays. (a) Top-view maximum amplitude projection (MAP), front-view MAP, and the cross-sectional slice along the green dashed line in the top-view MAP of the acoustic source. (b) Top-view MAP, front-view MAP, and corresponding cross-sectional slice along the green dashed line in the top-view MAP of the UBP reconstruction results with 505, 1009, 2006 sensors, respectively. (c) Top-view MAP, front-view MAP, and corresponding cross-sectional slice along the green dashed line in the top-view MAP of the SlingBAG Pro reconstruction results with 505, 1009, 2006 sensors, respectively. (d) Imaging setup. (e) Point cloud iteration process of SlingBAG Pro reconstruction with 2006 sensors. (Scale bar: 10 mm.)
  • Figure 2: Quantitative accuracy assessment of SlingBAG Pro reconstruction results. (a) Top-view maximum amplitude of Ground truth and SlingBAG Pro reconstruction results with different sensor numbers. (b) Amplitude comparison along the green line in top-view MAP of SlingBAG Pro reconstruction result.
  • Figure 3: Comparison of decay for both the ball numbers and the loss in coarse reconstruction stage between SlingBAG Pro and SlingBAG. (a-c) Decay of the ball numbers in coarse reconstruction stage for 505, 1009, 2006 sensor array, respectively. (d-f) Loss decay in coarse reconstruction stage for 505, 1009, 2006 sensor array, respectively.
  • Figure 4: Comparison of 3D photoacoustic reconstruction results under sparse irregular arrays. (a) Top-view maximum amplitude projection (MAP), front-view MAP, and the cross-sectional slice along the green dashed line in the top-view MAP for 3D volume of randomly initialization after zero-gradient filtering with 505, 1009, 2006 sensors, respectively. (b) Point cloud result and 3D volume of randomly initialization after zero-gradient filtering with 505, 1009, 2006 sensors, respectively. (c) Top-view MAP, front-view MAP, and corresponding cross-sectional slice along the green dashed line in the top-view MAP of the SlingBAG Pro reconstruction results with 505, 1009, 2006 sensors, respectively. (d) Point cloud result and 3D volume of SlingBAG Pro reconstruction with 505, 1009, 2006 sensors, respectively. (Scale bar: 10 mm.)
  • Figure 5: 3D PA reconstruction results of a rat liver. (a) 3D volume, XY Plane-MAP, XZ Plane-MAP, YZ Plane-MAP and the cross-section slice at green dashed line marked in XZ Plane-MAP of the SlingBAG Pro 3D reconstruction results using 1024 sensor signals. (b) XY Plane-MAP, XZ Plane-MAP, YZ Plane-MAP and the cross-section slice at green dashed line marked in XZ Plane-MAP of the UBP 3D reconstruction results using 1024 sensor signals. (Scale: 2 mm.) (c) Hierarchical optimization process of point cloud iterative reconstruction. (d) Schematic diagram of the imaging area.
  • ...and 2 more figures