Speed detection to suppress motion artifacts (MA) in laser speckle contrast imaging (LSCI)

Abstract

Laser speckle contrast imaging (LSCI) is an optical technique for noninvasive assessment of microcirculatory blood flow. LSCI has a broad application in medicine including dermatology. Since laser speckles are the basis for this imaging modality, any external motions during a measurement from both patient and operator affect the blood flow images. This challenge is called motion artefacts (MA). Here, we propose a complete procedure for analysis of speckles, that is, pre-segmentation, segmentation, motion detection, spatial alignment, perfusion map calculation and MA suppression. The handheld perfusion imager (HAPI) operated in both handheld and mounted schemes, has been used for measurements on 14 psoriasis subjects. The advantage of HAPI is use of a single monochromatic camera for both speckle imaging and motion detection. We make use of the black marker dots (made by the clinical investigator to determine visual psoriasis lesion boundary) for calculation of two-dimensional displacements of HAPI during each measurement (i.e. on-surface displacements). These on-surface displacements are integrated to translate each speckle image back to the initial position at the start of the measurement (i.e. spatial alignment). Furthermore, in handheld measurements, MA corrected blood flow maps (also called perfusion maps) are formed by extrapolation of a linear fit from local perfusion versus detected speed to the zero speed, that is, a value ideally always lower than the local mean perfusion. We show that our MA suppression technique makes handheld perfusion maps more similar to the associated mounted perfusion maps in term image histograms and mean values.