Published: Jan. 1, 2025
Improving the accuracy of nailfold capillary static parameter measurements is currently a popular area research. Currently, division areas usually based on experience clinical practitioner, and those with less exhibit high error rates. This study aims to reduce rate by developing an automated method for measuring parameters capillary. First, we used digital image processing combined deep learning accurately delineate that correspond apical, arterial, venous branches. Subsequently, calculated pipe diameters three limbs were compared manually measured diameters. The experimental results demonstrate RMSE (μm) arterial limb, limb values are 0.614, 0.71, 0.686, respectively. proposed 91%. has in measurement capillary, providing effective solution time-consuming process labor costs
Language: Английский