首页> 外国专利> METHOD FOR FORMING A two-dimensional image two-dimensional image biosignal AND ANALYSIS

METHOD FOR FORMING A two-dimensional image two-dimensional image biosignal AND ANALYSIS

机译:二维图像二维图像生物信号的形成方法及分析

摘要

1. A method for forming two-dimensional image and its analysis biosignal, comprising the formation of the phase plane of a dynamic system, characterized in that in the system selected two subsystems, the processes which occur in the form of quasi-synchronized signal x (t) and x (t), is performed synchronous recording these signals during a time T, determined by the nature and objectives of studies investigated system, at the coordinates x (t) -x (t) build phase portrait investigated dynamical system in the interval T, define h mknutuyu curve contouring the portrait, the Fourier descriptors determined closed curve obtained, after which the classification is carried out by means of a dynamic system investigated trainee classifier, constructed in the space of descriptors Fure.2. A method according to claim 1, characterized in that for training the classifier training set is formed, consisting of images of closed curves contouring phase portraits known classes of dynamic systems built in the coordinates x (t) -x (t), the Fourier descriptors determined all curves included in the training set, determine the number of descriptors kto each i-th training sample curve determined curve with a maximum number of descriptors Kdeskriptoram each i-th curve in the high frequency region is added with a K-value of zero Kdeskriptorov amplitudes of all spectral components in the spectrum of each i-th curve is multiplied by the value K / Kzadavshis error recovery curves on Fourier descriptors and an optimizing criterion, determine the number of descriptors K, used for the labeling of the dynamic system, and an
机译:1.一种形成二维图像及其分析生物信号的方法,包括形成动态系统的相平面,其特征在于,在系统中选择了两个子系统,这些过程以准同步信号x的形式发生。 (t)和x(t)在时间T内同步记录这些信号,时间T由研究系统的性质和目标确定,在坐标x(t)-x(t)上建立相画像研究的动力系统在间隔T内,定义轮廓肖像的h mknutuyu曲线,傅立叶描述符确定闭合曲线,然后通过动态系统研究的学员分类器在描述符Fure.2的空间中进行分类。 2.根据权利要求1所述的方法,其特征在于,形成用于训练的分类器训练集,其包括闭合曲线的图像,所述闭合曲线的轮廓描绘出相图,所述相图建立在坐标x(t)-x(t)中的已知动态系统,傅立叶描述符中。确定训练集中包含的所有曲线,确定每条第i条训练样本曲线的描述符数k确定的描述符具有最大描述符数Kdeskriptoram的曲线高频区域中的每条第i条曲线的K值均为零每条第i条曲线的频谱中所有频谱分量的Kdeskriptorov振幅乘以傅立叶描述符上的值K / Kzadavshis误差恢复曲线并确定一个优化标准,确定用于动态系统标记的描述符数量K,和

相似文献

  • 专利
  • 外文文献
  • 中文文献
获取专利

客服邮箱:kefu@zhangqiaokeyan.com

京公网安备:11010802029741号 ICP备案号:京ICP备15016152号-6 六维联合信息科技 (北京) 有限公司©版权所有
  • 客服微信

  • 服务号