Speaker identification based on classify feature sub - space gaussian mixture model and neural net fusion 基于分类特征空间高斯混合模型和神经网络融合的说话人识别
In the fourth chapter we deal with the eigenspace ( see 4 . 1 ) and singular space ( see 4 . 2 ) perturbations 第四章讨论了特征空间(见4 1 )和奇异空间(见4 2 ]的扰动。
If the feature differences are small and even cross in the original space , it is difficult for traditional algorithms to cluster correctly 引入核函数,将原始数据由数据空间映射到特征空间,在特征空间中进行聚类。
This paper presents a method of analyzing stocks by fuzzy core clustering in character space formed by techno - index 摘要提出采用技术指标构造特征空间,在特征空间上用模糊核聚类算法寻找股市规律的股市技术分析方法。
This two kinds of methods make a good use of the information supplied by multiple - dimensional feature space and enhance the accuracy of segmentation 这两种方式有效地利用了多维特征空间对于分割目标所提供的信息,使分割的准确性有了较大的提高。
This is a two - objective optimization problem . to solve the problem , two methods are presented , which are the weighted method and the eigenspace method 文中还给出了两种确定权值的方法,即加权法和特征空间法,并通过计算机仿真对这两种方法进行了比较。
In the field of human face tracking , we contrast several main similarity methods nowadays , and adopt a method of human face tracking based on feature space model 在人脸跟踪领域,本文对比了当前几种主要的相似度匹配方法,采用了一种基于特征空间模型的人脸跟踪方法。
2 ) using the multi - character of the wavelet space as the characters of a target to increase the dimension of the character space of the target and to increase the reliability of identifying the target 利用小波空间的多种特征作为目标特征,提高特征空间的维数,增加目标识别的可靠性。
4 . to solve the pattern recognition problems , the neural network can achieve the involved demarcate of earmark space , and it cut out for high acceleration and parallel systems to achieve 4 、神经网络对于解决模式识别问题来说,可以实现特征空间较复杂的划分,适合于高速并行处理系统来实现。
Fisher rule is used then to find the best line that suitable for classification , according to which the sample data is projected from a high - dimensional space to a low - dimensional space 该方法利用fisher判别准则,找到一条最好和最易分类的投影直线,并将样本数据由高维的图像空间投影到低维的特征空间。