Laser speckle auto-inverse covariance imaging for mean-invariant estimation of blood flow.

Abstract

Laser speckle contrast imaging maps the changes in blood flow by estimating the decorrelation time of a scattered light field. However, speckle contrast is a biased statistics estimator that results in a theoretic bias between its expected value and the true value. Moreover, the average of speckle contrast depends on the statistical sampling size, which further hinders the estimation of decorrelation time from speckle contrast. Here, we present a new, to the best of our knowledge, unbiased statistics analysis based on auto-inverse covariance to improve the estimation of decorrelation time using laser speckle. Theoretical and experimental results demonstrate that the speckle auto-inverse covariance analysis is mean-invariant, so that the average of the estimation is not dependent on the sampling size. Furthermore, it can produce less statistical fluctuation, especially for slow flow, and consume less computation time than that of speckle contrast analysis.