Open AccessGeometrical capillary network analysis
Author(s) -
Sainthillier JeanMarie,
Degouy Arnaud,
Gharbi Tijani,
Pieralli Christian,
Humbert Philippe
Publication year - 2003
Publication title -
skin research and technology
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.521
H-Index - 69
eISSN - 1600-0846
pISSN - 0909-752X
DOI - 10.1034/j.1600-0846.2003.00037.x
Subject(s) - computer science , artifact (error) , microcirculation , computer vision , artificial intelligence , image processing , process (computing) , software , capillary action , image (mathematics) , materials science , medicine , composite material , radiology , programming language , operating system
Background: Skin microcirculation, especially the superficial network, can be assessed by a computer capillary video microscope system. The study of morphology and dynamics of microcirculation must include all dynamic and cooperative processes between the capillaries. For characterizing capillary ensembles, the statistical and geometrical properties of the network need to be explored. Methods: The microvaculature of the skin and the microcirculation were investigated by combining videocapillaroscopy (VCP) and image processing techniques based on computational geometry and graph theory. Our goal was to characterize the capillary network in noisy pictures of the scalp. Different geometric methods were developed, based on proximity parameters (distance and surface) in order to circumscribe and construct this network. Results: By studying the distribution of these parameters, extreme values or outliers, which usually correspond to artifact subregions in the pictures could be eliminated. Different algorithms were developed and has been implemented in an image processing software (Capilab Toolbox). Conclusion: This computerized system is capable of real‐time processings, increasing the quality of videocapillaroscope images and minimizing the disturbance of artifacts. The algorithms presented here are easy to implement and can process any kind of images of the skin, even in the scalp. In association with an example‐based detection system, this method can be generalized to other stimuli in the same conditions.