Title :
Based on an improved pre-PCA + LDA classifier design in tumor cells
Author :
Tu, Chunping ; Gan, Lan ; Yu, Zhongping
Author_Institution :
Sch. of Inf. Eng., East China Jiaotong Univ., Nanchang, China
Abstract :
Through the integration of global and local classifiers can be better to image normal and abnormal (including cancer and hyperplasia) of the classification, But early cancer and hyperplasia in shape have only a very slight difference. Therefore, we also need the new features to design a new classifier to realize the classification of cancer and hyperplasia. In this paper, we use the improved PCA + LDA dimension reduction method to overcome the traditional PCA + LDA method, using which the test samples have poor generalization ability. at the same time, We have done a related experiment to achieve the cancerous cells and proliferation of cells in the classification problem. compared With the traditional PCA + LDA classification methods, we found the improved PCA + LDA to achieve a better recognition effect.
Keywords :
cancer; image classification; medical image processing; principal component analysis; tumours; LDA classifier design; PCA; cancer; hyperplasia; tumor cells; Accuracy; Biomedical imaging; Cancer; Integrated circuits; Nickel; Tumors; Generalization; Global and local classifiers; PCA + LDA classification;
Conference_Titel :
Computer and Communication Technologies in Agriculture Engineering (CCTAE), 2010 International Conference On
Conference_Location :
Chengdu
Print_ISBN :
978-1-4244-6944-4
DOI :
10.1109/CCTAE.2010.5543500