Sparse coding-based denoising and contrast enhancement in low-fluence multi-parametric photoacoustic microscopy

We report a sparse coding-based two-stage image processing technique for noise reduction and contrast enhancement in photoacoustic microscopy (PAM). We have assessed the performance of this denoising technique using multi-parametric PAM images acquired at different optical fluence levels in vivo. Our results show that this technique can significantly improve the contrast-to-noise ratio (~7 times) and visibility of the microvasculature in low-fluence noisy images, enabling us to reduce the optical fluence by a factor of 5 while maintaining a similar image quality. Moreover, this technique improves the accuracy of quantitative measurements of blood oxygenation and flow in low-fluence PAM images.