Acoustic Resolution Photoacoustic Microscopy Imaging Enhancement: Integration of Group Sparsity with Deep Denoiser...

Area of Science:

• Medical Imaging
• Biomedical Engineering
• Image Reconstruction

Background:

• Acoustic resolution photoacoustic microscopy (AR-PAM) offers deep tissue imaging but suffers from degraded resolution due to acoustic focusing limitations.
• Existing model-based reconstruction methods struggle to capture fine anatomical details in AR-PAM images.

Purpose of the Study:

• To enhance AR-PAM imaging resolution and structural detail reconstruction.
• To mitigate artifacts introduced by traditional patch-based reconstruction methods.
• To leverage complementary information from external datasets and deep learning for improved image quality.

Main Methods:

• Proposed a novel group sparsity prior for simultaneous reconstruction, exploiting non-local structural similarity within AR-PAM images.
• Integrated an external image dataset and a deep denoiser prior with the group sparsity prior to reduce artifacts.
• Validated the method on simulated and in vivo AR-PAM data, comparing with optical resolution photoacoustic microscopy (OR-PAM) results.

Main Results:

• Significantly improved image quality in simulated AR-PAM data, with PSNR increasing from 16.36 to 27.62 dB and SSIM from 0.46 to 0.92.
• Achieved superior local and global perceptual qualities in in vivo AR-PAM images.
• Markedly increased SNR from 10.59 to 30.83 and CNR from 8.61 to 27.54 in in vivo results.

Conclusions:

• The proposed method effectively enhances AR-PAM image resolution and detail.
• Combining group sparsity with deep learning priors successfully mitigates artifacts and improves reconstruction quality.
• The advanced AR-PAM technique shows significant potential for broader medical and clinical applications.