【摘要】 针对图像的分类识别问题,首先,提出在目前尺度不变特征变换(SIFT)的基础上,融合完全随机树(ERT)构建新的特征进行图像分类识别的方法;其次,在ERT生长过程中,将现有的绝对熵计算模式改为香农熵(相对熵)模式,增强了ERT的稳定性,使得ERT的适应性更好;最后,对任意选择的分裂条件进行了最小熵(Smin)限制,使得融合了SIFT的ERT在不失随机性的基础上,有更好的聚类效果,最终使分类识别方法稳定性更好,精度更高.更多还原

【Abstract】 There are two contributions of this research to the image classification and recognition.Firstly,it changes the existing computing model of absolute entropy to the relative entropy(Shannon entropy) mode,increasing the stability of ERT(extremely randomized tree) and ERT’s adaptability when generating the ERT.On the basis of the existing methods of using SIFT(scale invariant feature transform),ERT and SVM(support vector machine) are added in image classification and recognition.Secondly,it carries a minimum entropy limit in the choice of randomized selecting splitting conditions on the basis of not losing the randomness,obtaining better clustering results and thus making the method get better effect in classification and recognition with higher accuracy.更多还原