| Image thinning algorithm(also called skeletonizing algorithm)can keep the image original topology and remove edge pixels,until the processed image as a single pixel wide.It is an important part of image preprocessing.A good image thinning method is beneficial to the feature extraction of the image and the improvement of the accuracy rate of image recognition.The method is significant to the performance of the whole image recognition system.Spiking neural P systems(SN P systems,for short)are a computing model under the framework of membrane computing and belong to the third generation neural network.The computing power is equivalent with Turing machine.The research of Spiking neural P systems mainly focuse on theoretical research.How to apply it for image processing has risen many researchers' interest.In this paper,we study spiking neural P systems and their applications in image processing.In view of the uncertainty of the calculation of spiking neural P systems,spiking neural P systems with the priority of spiking rules is proposed in this paper.And a parallel image thinning method based on spiking neural P systems is proposed.The formal definition,topological structure and introduce are given.Finally,the experimental simulation results show that the image thinning method in this paper has higher data compression efficiency.Under the ensurence of the original image topology,the skeleton of the image is simpler and the skeleton burr is less.For the thinning of handwritten character images,the burr of the image skeleton obtained by this method is much less than that obtained from the Spiking neural P systems based on Guo and Hall algorithm.In addition,this method can achieve effective thinning for fingerprint image and guidepost image.Therefore,the method proposed in this paper can realize image data compression and can be applied to image recognition preprocessing such as character recognition,guidepost recognition and fingerprint recognition. |
PDF Full Text Request