Name
Abstract | ||
|---|---|---|
A family of linear subspace classifiers called local subspace classifier (LSC) outperforms the k-nearest neighbor rule (kNN) and conventional subspace classifiers in handwritten digit classification. However, LSC suffers very high sensitivity to image transformations because it uses projection and the Euclidean distances for classification. In this paper, I present a combination of a local subspace classifier (LSC) and a tangent distance (TD) for improving accuracy of handwritten digit recognition. In this classification rule, we can deal with transform-invariance easily because we are able to use tangent vectors for approximation of transformations. However, we cannot use tangent vectors in other type of images such as color images. Hence, kernel LSC (KLSC) is proposed for incorporating transform-invariance into LSC via kernel mapping. The performance of the proposed methods is verified with the experiments on handwritten digit and color image classification. |
Year | DOI | Venue |
|---|---|---|
2008 | 10.1093/ietisy/e91-d.6.1756 | IEICE Transactions |
Keywords | Field | DocType |
handwritten digit,local subspace classifier,handwritten digit recognition,tangent vector,conventional subspace classifier,linear subspace classifier,image classification,handwritten digit classification,color image classification,kernel lsc,classification rule,k nearest neighbor,color image,euclidean distance | Kernel (linear algebra),Classification rule,Pattern recognition,Subspace topology,Computer science,Random subspace method,Tangent vector,Artificial intelligence,Kernel method,Contextual image classification,Color image | Journal |
Volume | Issue | ISSN |
E91-D | 6 | 1745-1361 |
Citations | PageRank | References |
0 | 0.34 | 12 |
Authors | ||
1 | ||
Name | Order | Citations | PageRank |
|---|---|---|---|
| Seiji Hotta | 1 | 6 | 4.98 |