Research On Extraction Methods Of Photoacoustic Microvascular Images

Optical resolution photon microscopy(OR-PAM)is an imaging method based on photoacoustic effect,which combines the advantages of high resolution of optical imaging and considerable depth of ultrasound imaging together.Through using endogenous hemoglobin as control,OR-PAM system can image the morphology and function of microvascular in vivo with high-resolution which could reach micron / sub-micron level,thus make it could take image for a single capillary.Therefore,OR-PAM is becoming a powerful tool of microcirculation study of the physiological or pathological related tissues.However,when comes to blood vessels tiny or deep,due to the light scattering and less light absorption of tissue etc.,OR-PAM has a poor signal-to-noise ratio,leads to vagueness signal and discontinuous image in weak signal regions.Therefore,it is necessary to use image post-processing technology to enhance these vessels images,improve the accuracy and quality of them,and lay the foundation for subsequent vascular information extraction and analysis.Many researchers have carried out vascular image processing and analysis with the traditional imaging techniques,such as X-ray imaging,optical coherence tomography(OCT),nuclear magnetic resonance imaging(NMR),and CT imaging.Recently,there are also relevant researches on the photoacoustic vascular image through OR-PAM method.However,these studies are related with the common blood vessels,not the sophisticated tumor microvessels.Analysis of the complicated microvascular images obtained from the OR-PAM system with high precision calls for a more effective method of image preprocessing and segmentation.Therefore,in this paper,an effective microvessel segmentation algorithm based on high frequency emphasis filtering,multi-scale Hessian matrix and gradient domain high dynamic range compression is proposed to enhance the microvascular image and segmentation of OR-PAM system.The results of the study carried in healthy mices through OR-PAM system have verified the ability of OR-PAM on accurately subdivision and quantification for blood vessels,including for capillary vessel.Moreover,in order to objectively evaluate the growth and progression of tumor microvessels through OR-PAM system,we introduced four parameters which have been widely accepted to describe and evaluate vascular morphology: vascular diameter,vascular density,vascular curvature and fractal dimension.In order to verify the proposed method in this study,we performed normal OR-PAM imaging both for normal blood vessel and murine melanoma of the same ear of mouse,and obtained the photoacoustic image of the normal blood vessels,melanoma and the peripheraltissues at different time points.Then the relevant images were processed,and quantified with the four parameters values.The comparing results finally showed that our method could effectively enhance the photoacoustic blood vessel image,extract the normal and tumor microvessels information with high precision,and quantitatively evaluate the microvessels from various aspects.The results also elaborated the growth,invasion and metastasis of the microvessels of melanoma tissue,and it provides an objective pathological reference for clinicians and physicians on tumor estimation,diagnosis,treatment and monitoring.