Automatic Image Segmentation Algorithm Based on PCNN and Fuzzy Mutual Information

The pulse coupled neural network (PCNN) algorithm has been effectively used in image segmentation. In this paper, we proposed a new image auto-segmentation algorithm based on PCNN and fuzzy mutual information (FMI). The image was firstly segmented by PCNN, and then FMI was used as the optimization criterion to automatically stop the segmentation with the optimal result. Different images were segmented by max-FMI PCNN, Otsu segmentation algorithm and max-entropy PCNN to evaluate the segmentation accuracy. The experimental results demonstrated that the CT and ultrasound images could be well segmented by the proposed algorithm with strong robustness against noise. The results suggest that the proposed algorithm can be used for medical image segmentation.

[1]  N. Otsu.  A threshold selection method from gray level histograms , 1979 .

[2]  Reinhard Eckhorn,et al.  Feature Linking via Synchronization among Distributed Assemblies: Simulations of Results from Cat Visual Cortex , 1990, Neural Computation.

[3]  R. Eckhorn,et al.  High frequency (60-90 Hz) oscillations in primary visual cortex of awake monkey. , 1993, Neuroreport.

[4]  Dimitris N. Metaxas,et al.  Combining low-, high-level and empirical domain knowledge for automated segmentation of ultrasonic breast lesions , 2003, IEEE Transactions on Medical Imaging.

[5]  Liu Qing,et al.  Automated image segmentation using improved PCNN model based on cross-entropy , 2004, Proceedings of 2004 International Symposium on Intelligent Multimedia, Video and Speech Processing, 2004..

[6]  Robert Rohling,et al.  Ultrasound image segmentation using spectral clustering. , 2005, Ultrasound in medicine & biology.

[7]  Nikos Dimitropoulos,et al.  A hybrid multi-scale model for thyroid nodule boundary detection on ultrasound images , 2006, Comput. Methods Programs Biomed..

[8]  Wei Cai,et al.  A region-based multi-sensor image fusion scheme using pulse-coupled neural network , 2006, Pattern Recognit. Lett..

[9]  Liu Qing,et al.  Image Target Recognition Using Pulse Coupled Neural Networks Time Matrix , 2006, 2007 Chinese Control Conference.

[10]  J. Alison Noble,et al.  Ultrasound image segmentation: a survey , 2006, IEEE Transactions on Medical Imaging.

[11]  Yide Ma,et al.  A Novel Algorithm of Image Enhancement Based on Pulse Coupled Neural Network Time Matrix and Rough Set , 2007, Fourth International Conference on Fuzzy Systems and Knowledge Discovery (FSKD 2007).

[12]  Yide Ma,et al.  Pathologic Region Detection Algorithm for Prostate Ultrasonic Image Based on PCNN , 2007, FAW.

[13]  Wu Cheng-mao.  Fuzzy Mutual Information and Its Application in Image Segmentation , 2008 .

[14]  D. Arar,et al.  A fast technique for gray level image thresholding and quantization based on the entropy maximization , 2008, 2008 5th International Multi-Conference on Systems, Signals and Devices.

[15]  Zhang Yi,et al.  A mixed noise image filtering method using weighted-linking PCNNs , 2008, Neurocomputing.