Multi-focus Image Fusion via Blind Deconvolution and Registration based on Photoacoustic Microscopy

Abstract

Abstract Optical resolution photoacoustic microscopy (OR-PAM), which can present the microstructure of biological tissues is useful in the disease diagnosis. However, conventional OR-PAM can only obtain images with high resolution in a narrow range of depths around the focal plane. In this study, we proposed a multi-focus photoacoustic image fusion method based on the blind deconvolution. This method is divided into two steps. First, we used the method of blind deconvolution to obtain an approximate point spread function (PSF) for each photoacoustic images obtained from multiple layers, and then got the corresponding deblurred photoacoustic images. Second, we used the method of multi-focus images registration and fusion to accurately align the multi-focus photoacoustic images and effectively fuse the features of different images. Finally, the proposed method was applied to process the photoacoustic images obtained from photoacoustic imaging experiments on ants with irregular morphology using fast controllable confocal focus PAM (FC-PAM). The results show that our method improves the quality of photoacoustic images, and can quickly obtain multi-focus fused photoacoustic images with high resolution. It indicates that the proposed multi-focus photoacoustic image fusion method based on FC-PAM has great potential in fast imaging of living biological tissues at multiple imaging depths.