Chaotic Neurons Network with Plastic Couplings for Image Segmentation and Posterization

本研究的目的是开发一种适用于高分辨率、连续色调图像的动态图像分割系统的基本技术。基于非线性动力系统理论，分析了具有塑性耦合的离散时间混沌神经元的动力学行为。研究结果为动态图像分割提供了合适的系统参数取值。该系统具有可调参数值，适用于多幅医学图像。此外，在图形处理单元中实现所提出的系统比在中央处理单元中执行的动态图像分割产生更快的动态图像分割。

The purpose of this study is to develop basic technologies on a dynamic image-segmentation system that is applicable to high-resolution, continuous-tone images. On the basis of nonlinear dynamical systems theory, we analyzed the dynamics of discrete-time chaotic neurons with plastic couplings. The results suggested appropriate values of system parameters for dynamic image segmentation. The proposed system with tuned parameter values worked for several medical images. Moreover, an implementation of the proposed system into a graphics processing unit yielded faster dynamic image segmentation than that executed in a central processing unit.