Abstract
Diffuse speckle contrast analysis (DSCA) is a valuable technique for monitoring blood flow speed, but its accuracy and sensitivity are often compromised by system noise and non-ergodicity, particularly in deep tissue measurements. Noise and static scattering tissues lead to overestimated speckle contrast and consequent underestimation of blood flow speed, especially in regions of rapid flow. To address these challenges, we developed a noise correction method integrated with a non-ergodicity calibrated model to enhance the sensitivity and accuracy of DSCA. A guide for selecting an appropriate core diameter and numerical aperture was provided for multi-mode fiber DSCA. Validation through phantom and in vivo experiments demonstrated that the non-ergodicity and noise-corrected DSCA improved the sensitivity of blood flow speed measurements in deep tissues.