| From the 90s of the 20th century, through the study of cat visual cortex neurons pulse synchronization concussion, Eckhorn got the model of mammalian neurons. By a few improvements on Eckhorn's model, we got the Pulse-Coupled Neural Networks model. This model has tow characteristics: two-dimensional space similar to the image and pixel gray grouping similar characteristics, and it can reduce the local gray image margin, partially offset images intermittent small, it is the other image segmentation method unmatched features. PCNN is mainly used for feature extraction, the edge information analysis, image segmentation, target identification. PCNN currently based on the model of large-scale diagnostic imaging systems, military target recognition systems, image segmentation and target classification system are being developed.This paper applies PCNN to medical image processing, in the HIS color space. The algorithm can smooth image, improve the image edge, and at the same time, improve the medical image visual effects and color image of the true effect by balanced handling the brightness intensity and nonlinear exponential adjust of the saturation. At first, this paper gives a brief overview on the development of PCNN and development of image enhancement. On this basis, the paper gives the system's hardware and software details. Hardware is Blackmagic company's Multibridge video capture equipment, through PCI-E bus, it can achieve high-speed video acquisition. At the same time a video collection and a D/A or A/D converter functions are available. On software aspect, use MATLAB to achieve PCNN simulation, and through unique advantages of PCNN to enhance the medical image processing. First convert RGB image to HIS image, and then process I and S to get better effect, and finally, convert HIS image to RGB image.After debugging, system can work normally and can be connected directly to a monitor. And system provides a good interface and expansibility for latter part. |
PDF Full Text Request