Self-Supervised Speed of Sound Recovery for Aberration-Corrected Photoacoustic Computed Tomography

2024-09-17

AI Summary

In photoacoustic computed tomography (PACT), spatial heterogeneity in the speed of sound (SOS) induces wavefront distortion, severely degrading image quality. Direct SOS measurement is experimentally cumbersome and computationally expensive, while supervised learning approaches are hindered by scarce annotated SOS data. This paper proposes a differentiable physics-informed, self-supervised joint reconstruction framework that jointly optimizes SOS distribution and photoacoustic image without requiring ground-truth SOS labels. Its key innovations include: (1) the first semi-blind, end-to-end self-supervised SOS inversion framework; (2) dual parameterization—supporting both pixel-grid and neural field representations of SOS; and (3) integration of gradient-driven SOS updates with joint optimization. Experiments demonstrate quantitative superiority over baselines in simulations and significant aberration suppression in vivo, yielding marked improvements in spatial resolution and contrast-to-noise ratio, alongside a 35× speedup in inference time.

Technology Category

• Search and Optimization: Learning to Search
• Computer Vision: Learning & Optimization for CV
• Planning, Routing, and Scheduling: Optimization of Spatio-temporal Systems

Application Category

• Systems and Infrastructure for Web, Mobile and WoT: Applied ML and AI for Web-based mobile applications
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• User Modeling, Personalization and Recommendation: On-Device user modeling, personalization, and recommendation

Abstract

Photoacoustic computed tomography (PACT) is a non-invasive imaging modality, similar to ultrasound, with wide-ranging medical applications. Conventional PACT images are degraded by wavefront distortion caused by the heterogeneous speed of sound (SOS) in tissue. Accounting for these effects can improve image quality and provide medically useful information, but measuring the SOS directly is burdensome and the existing joint reconstruction method is computationally expensive. Traditional supervised learning techniques are currently inaccessible in this data-starved domain. In this work, we introduce an efficient, self-supervised joint reconstruction method that recovers SOS and high-quality images using a differentiable physics model to solve the semi-blind inverse problem. The SOS, parametrized by either a pixel grid or a neural field (NF), is updated directly by backpropagation. Our method removes SOS aberrations more accurately and 35x faster than the current SOTA. We demonstrate the success of our method quantitatively in simulation and qualitatively on experimentally-collected and in-vivo data.

Problem

• Improves image quality in photoacoustic computed tomography
• Recovers speed of sound in tissue efficiently
• Reduces computational cost of joint reconstruction

Innovation

• Self-supervised joint reconstruction method
• Differentiable physics model for SOS recovery
• Neural field parametrization for SOS updates