Efficient temporal pattern recognition by means of dissimilarity space embedding with discriminative prototypes

详细信息    查看全文

• 作者：Brian Kenji Iwana^a ; ^b ; ^{brian@human.ait.kyushu-u.ac.jp}Author Vitae ; Volkmar Frinken^a ; ^c ; ^dAuthor Vitae ; Kaspar Riesen^e ; ^{kaspar.riesen@fhnw.ch}Author Vitae ; Seiichi Uchida^a ; ^{uchida@ait.kyushu-u.ac.jp}Author Vitae
• 关键词：Temporal patterns ; Online digit classification ; Dissimilarity representation ; Ensemble classification ; Dissimilarity space embedding
• 刊名：Pattern Recognition
• 出版年：2017
• 出版时间：April 2017
• 年：2017
• 卷：64
• 期：Complete
• 页码：268-276
• 全文大小：1771 K
• 卷排序：64

文摘

We propose a method of using dissimilarity space embedding for temporal patterns. Ensemble classification was used for prototype selection to increase the efficiency. We evaluate the performance of our method on online handwritten digits. The experiments showed a high accuracy with a very small number of patterns.