Cellular nonlinear network methodology and related technology was used for the implementation of a real-time image processing system for the study of structural and functional parameters in the microcirculation. The observation of these parameters is basic in the description and characterization of the physiological phenomena occurring in the peripheral vascular network at a micrometric scale. They contribute to the understanding of the global cardiovascular regulatory system for both experimental and clinical applications. The aim of this new CNN-based approach is to implement a real-time system that is based on automated image processing algorithms that overcome the limits of the conventional invasive, manual, and operator dependant methods. It would also allow an objective protocol in the in vivo experiments. These conditions are necessary in biological studies in order to make the experimental results reproducible and comparable. An algorithm was implemented, by exploiting the CNN structure potentiality, in order to characterize the morphology of the capillary maps, to determine the red blood cells density in blood (Hematocrit), and to calculate red blood cells velocity (RBCV) in capillaries from image sequences captured during in vivo experiments by intravital microscopy. The algorithm was designed and tested using a CNN simulator and then optimized and implemented via hardware on the ACE16k CNN chip. The final results of the image processing were furthermore compared with measurements obtained with conventional, manual methods used in microvascular studies.
|Titolo:||Cellular Nonlinear Network in microcirculation characterization|
|Data di pubblicazione:||2006|
|Appare nelle tipologie:||1.1 Articolo in rivista|