Signal-to-noise Ratio Improvement of Laser Diode Acoustic Resolution Photoacoustic Microscopy

Acoustic Resolution Photoacoustic Microscopy (AR-PAM) can be used to achieve non-invasive imaging of subcutaneous blood vessels. Laser diode (LD) as an excitation source has become a trend to promote the clinical application of PAM. However, the LD-AR-PAM system usually has a low SNR problem due to low energy and electrical noise interference. In this paper, firstly, based on the temporal synchronous assumption of B-scan data, singular value decomposition (SVD) has been applied for electrical decluttering in B-scan data. Then, for high quality and low computational cost image reconstruction, a correlation weighting factor is proposed to support the Synthetic Aperture Focusing Technique (SAFT) imaging algorithm. The proposed Sum Mean Standard Deviation (SMSD) weighting factor can use positive and negative values to directly characterize sound sources and artifacts. It differs significantly from conventional weighting, which only suppresses artifacts rather than eliminating them, such as the Coherence Factor (CF). The LD-AR-PAM system with the above method has been applied to human hair image reconstruction. SVD decluttering results in an average SNR of 21.4 dB and an average improvement of 27.1%. For the maximum amplitude projection (MAP) image of a human hair, the lateral resolution of SAFT, SAFT+CF reaches 163.3 and 153.8μm. The lateral resolution of the SAFT+SMSD surpasses that of the above beamformers, reaching 96.8μm. The corresponding SNRs are 23.9, 26.2 and 29.2dB respectively. The final SNR is improved by more than 37% compared to the source hair image. The results prove that our SNR improvement method greatly improves the quality of LD-AR-PAM imaging, which is an important proof of concept for the clinical application of this technology.