Superior perturbation inversion strategy based on Markov random field incorporating measurement matrix optimization in linear array photoacoustic tomography

Photoacoustic tomography technology, which combines the advantages of deep tissue penetration of ultrasound imaging and high contrast of optical imaging, has emerged as a prominent research area in biomedicine. However, the accuracy of its imaging is inherently constrained by factors such as detection angle and energy attenuation in the medium, presenting ongoing challenges in enhancing efficiency. In response to these challenges, we propose a superior perturbation inversion strategy based on the Markov random field incorporating measurement matrix optimization coupled with Linear-Extremum-guided interpolation (PMMLI). This model leverages spatiotemporal information captured during the linear array detection process to enhance signal continuity. Moreover, the iterative results are guided toward a superior pathway by optimizing the measurement matrix and incorporating the Markov Random Field prior information of diagonal neighbors as a supplementary perturbation criterion in reconstruction. Through imaging of semi-transparent media containing vascular-shaped absorbers, our results demonstrate that PMMLI yields smoother background images and more detailed vascular structures compared to alternative methods (the conjugate gradient method, the piecewise cubic Hermite interpolation method, and Linear-Extremum-guided interpolation method). The Structural Similarity Index Measure (0.999) and Peak Signal Noise Ratio (59.997 dB) of the PMMLI model are three times and two times higher than those of other methods (about 0.3 and 20 dB), respectively. The Root Mean Square Error is only 0.094 %, which is two orders of magnitude lower compared to other technologies. Notably, even in scenarios where up to 30 % measurement error is introduced, PMMLI maintains the ability to discern main vascular structures, underscoring its robustness. This research lays a foundation for advancing the accuracy and reliability of photoacoustic tomography, showcasing the potential of PMMLI in addressing existing limitations and enhancing imaging outcomes in biomedicine.