ABSTRACT
Histogram-based thresholding techniques utilized for cell-free layer width measurement in arteriolar flow may produce an overestimation of the layer width since they do not consider faint shaded regions near the vessel wall as part of the erythrocyte column. To address this problem, we developed a new method for detecting the boundary of the erythrocyte column based on an edge detection algorithm. This automated method (grayscale method) provides local detections of the inner vessel wall as well as the boundary between the cell-free layer and the erythrocyte column without binarization of grayscale images. The cell-free layer width measurements using the grayscale method and existing techniques (minimum method and Otsu's method) were compared with those determined manually in arteriolar flows of the rat cremaster muscle. In the absence of the shaded regions, values obtained by the grayscale method and minimum method were statistically in good agreement with the manual method but not in the case of Otsu's method. When the faint shaded regions were present, the grayscale method appeared to produce more accurate results than the minimum method and Otsu's method.
