- Issue:
- 1999年06期

- Page:
- 481-485

- Research Field:

- Publishing date:

- Title:
- A Novel Method of Grouped Decision and Its Application

- Author(s):
- HuangXiuwu YangJingyu GuoYuefei JingXiaoyuan
- School of Information,NUST,Nanjing 210094

- PACS:
- O235

- DOI:
- -

- Abstract:
- T his paper analyses the generalized opt imal set of discriminant vectors, present s a g rouped decision method based on generalized opt imal set of discriminant vectors for linear feature ext raction. We divide the w hole sample to several groups, and illust rate that the every g roup. s value of the Fisher discriminant function is higher than that of w hole sample theoret-i cally , so, every group’ s recognit ion correct rate is much higher than w hole correct rate. In order to test the ef ficiency of our method, w e apply it to the human facial recognit ion. Ex per-i mental results have show n: by using the same number of the generalized opt imal discriminant vectors, our method can get higher recognition correct rate than that not grouping. If we use the g rouped decision method , we can reach good recognit ion correct rate w ith very few generalized opt imal discriminant vectors, yet the ex ist ing methods must use many generalized optimal discriminant vectors to reach.

1 Foley D H, Sammon J W J. An optimal set of discriminant vectors. IEEE Trans on Computer, 24( 3) : 281~ 289

2 Hong Z Q. Alg ebraic feature ex traction of image for recognition. Pattern Recognition, 1991, 24( 3) : 211~ 219

3 洪子泉, 杨静宇. 用于图像识别的图像代数特征抽取.自动化学报, 1992, 18( 2) : 233~ 237

4 Ho ng Z Q, Yang J Y. Optimal discriminant plane for a small number of samples and desig n method of classifier on t he plane. Pattern Recognition, 1991, 24( 3) : 317~ 324

5 Cheng Y Q, Zhuang Y M, Yang J Y. Optimal Fisher discr iminant analysis using the rank decompsition.Pattern Recog nit ion, 1992, 25( 1) : 101~ 111

6 Liu K, Cheng Y Q, Yang J Y. A generalized optimal set of discr iminant vectors. Pattern Recognition, 1992, 25( 7) : 731~ 739

- Memo:
- -

Last Update: 2013-03-29