Deep Learning-Enhanced Fiber-Coupled Laser Diode-Based Photoacoustic Microscopy for High-Resolution Microvasculature Imaging

Blood vessel imaging provides crucial information for physiological monitoring and disease diagnosis. Optical-resolution photoacoustic microscopy (PAM) is gaining intense attention in high-resolution in vivo imaging of blood vessels. However, its applicability in clinical settings is hindered by the requirement of bulky and costly pulsed lasers as well as complex alignment procedures. Here, as a novel approach to high-quality and cost-effective microvasculature imaging, a dual-wavelength fiber-coupled laser diode-based PAM (FC-LD-PAM) system is demonstrated on the mouse ear and brain in vivo. Furthermore, assisted with a deep learning method, the proposed technique effectively combines the advantages when using both small and large core-size multimode fibers-successfully enhances the resolution of blurry mouse microvascular images obtained using a large core-size fiber, revealing fine details that are comparable to those achieved using a small core-size fiber; significantly reduces the overall time for data acquisition up to fourfold, improving the efficiency significantly in the imaging process. In addition, the proposed approach can achieve accurate and high-resolution mapping of oxygen saturation in vivo, providing functional insight on living tissue. Therefore, the FC-LD-PAM can serve as a translational tool for high-resolution imaging with enhanced accessibility and versatility in the biomedical field. This study presents a dual-wavelength fiber-coupled laser diode-based photoacoustic microscopy (FC-LD-PAM) system that enables high-resolution in vivo blood vessel imaging and oxygen saturation measurement. With the assistance of a deep learning method, the proposed FC-LD-PAM system shows high image resolution, short acquisition time, and provides accurate oxygen saturation mapping, making it a versatile and accessible tool for biomedical applications. image