Model-Based Image Reconstruction for Linear Array Optoacoustic Imaging

Optoacoustic imaging is a promising biomedical imaging technique that combines optical contrast with ultrasonic resolution by detecting acoustic waves generated from pulsed light absorption in tissues. Two main issues with optoacoustic imaging are image reconstruction algorithms and the impact of probe geometry. Traditional reconstruction methods offer reasonable image quality but struggle with physical limitations; hence model-based (MB) reconstruction methods that incorporate complex tissue properties have been explored. Linear ultrasound probes are affordable and easy to manufacture, but they suffer from limited-view artifacts that degrade image quality. Concave probes, with better angular coverage, provide higher image quality but are more expensive and complex to manufacture. This study compares MB reconstruction using linear and concave probes across in-silico, in-vitro, and in-vivo images. The results show that regularized MB reconstruction can significantly enhance the quality of images obtained with linear probes, reducing the performance gap with concave arrays. For Shearlet L1 regularization, the structural similarity index (SSIM) was 0.66 ± 0.06, and the mean absolute error (MAE) was 0.057 ± 0.017, and a qualitative analysis revealed fewer artifacts in MB-reconstructed images when compared to traditional methods. Future work will focus on larger datasets and exploring deep learning to further improve MB reconstruction for linear arrays.