Automated quantitation of non-steady flow and lumen area based on temporal correlation

A robust method for automated quantitation of nonsteady flow and lumen area is presented. The method utilizes temporal information to differentiate between pixels within the lumen and a background region. Cross correlation (CC) is applied to identify the pixels with temporal information similar to a reference obtained from a pixel within the lumen region. The new method incorporates a computation of an optimal threshold; therefore, it provides an unbiased observer-independent quantitation. We have conducted in vitro and in vivo experiments to evaluate the reproducibility and accuracy of the method. Validation of the accuracy of the automated method was performed only on a phantom study since there is no other method that can provide the true lumen size measurement in vivo. The reproducibility of the automated method was evaluated using MRI data from subjects.