Pulse Coupled Neural Networks In Image Processing Applications

Pulse coupled neural network (PCNN), regarded as the third generation neural network, had been wildly used in image processing and path optimization problems. This network is proposed based on the study of mammal visual systems, and very similar to biological human neural network. Since this network is recently developed, it is essential to put this study further. The objective of this dissertation is to extend the applications of this network in the fields of image processing and pattern recognition.The main work is listed as follows:1.Reviewing the background, existed work and information processing advantages of PCNNs.2.Proposing a binary image thinning algorithm by using the autowaves of PCNNs. The main content of this algorithm is: the model of PCNN, parameter settings, autowaves meeting condition and evaluation methods.3.Investigating seal image segmentation methods, and applying PCNN and morphology to seal imprint extraction with complex background. The main work of this method is: holes filling methods, edge extraction and skeletization methods with selectivity the location of outer frame and seal imprint extraction by using the color information of outer frame.4.Based on the fact that diatom having closed membrance, applying PCNN to the extraction of skeleton and edge of diatom membrane. This method consists of two stags: extracting the skeleton of membrane and obtaining the edge of membrane based on the location of skeleton.5.Presenting a new code segmentation method. In this method the skeleton of code frame is first extracted, then by using the projection of this skeleton to X and Y axis, the location of every code is obtained.6.Introducing the Intelligent Chinese Seal Identification System. Background of seal identification is first reviewed, then the procedure of each processing steps is particularly investigated.