In vivo laser speckle imaging by adaptive contrast computation for microvasculature assessment

Abstract

Interference of light backscattered from a diffused surface leads to speckle formation in laser speckle imaging. These time integrated speckle patterns can be statistically analyzed to study the flow profile of moving scatterers. Simple speckle contrast analysis techniques have limited ability to distinguish thin structures due to presence of corrupting speckles. This paper presents a high resolution imaging technique by adaptive computation of contrast for Laser Speckle Contrast Analysis (adLASCA). Speckle images of retinal microvasculature in mice model are acquired during normal and reduced blood flow conditions. Initially, the speckle images are registered to compensate for movements, associated with heart beating and respiration. Adaptive computation is performed using local image statistics, estimated within a spatially moving window over successive time frames. Experimental evidence suggests that adLASCA outperforms other contrast analysis methods, substantiating significant improvement in contrast resolution. Fine vessels can be distinguished more efficiently with reduced fluctuations in contrast level. Quantitative performance of adLASCA is evaluated by computing standard deviation, corresponding to speckle fluctuations due to unwanted speckles. There is a significant reduction in standard deviation compared to other methods. Therefore, adLASCA can be used for enhancing microvasculature in high resolution perfusion imaging with reduced effect of corrupting speckles for effective assessment.