Incoherent Dictionary Pair Learning: Application to a Novel Open-Source Database of Chinese Numbers

被引：13

作者：

Abolghasemi, Vahid ^{[1
]}

Chen, Mingyang ^{[2
]}

Alameer, Ali ^{[3
]}

Ferdowsi, Saideh ^{[1
]}

Chambers, Jonathon ^{[3
,4
]}

Nazarpour, Kianoush ^{[3
]}

机构：

[1] Shahrood Univ Technol, Fac Elect Engn & Robot, Shahrood 36155316, Iran

[2] Univ Surrey, Dept Elect Engn, Guildford GU2 7XH, Surrey, England

[3] Newcastle Univ, Sch Engn, Newcastle Upon Tyne NE1 7RU, Tyne & Wear, England

[4] Harbin Engn Univ, Coll Automat, Harbin 0086, Heilongjiang, Peoples R China

来源：

IEEE SIGNAL PROCESSING LETTERS | 2018年 / 25卷 / 04期

基金：

英国工程与自然科学研究理事会;

关键词：

Chinese numbers; classification; incoherent dictionary pair learning; RECOGNITION; DESIGN; MODEL;

D O I：

10.1109/LSP.2018.2798406

中图分类号：

TM [电工技术]; TN [电子技术、通信技术];

学科分类号：

0808 ; 0809 ;

摘要：

We enhance the efficacy of an existing dictionary pair learning algorithm by adding a dictionary incoherence penalty term. After presenting an alternating minimization solution, we apply the proposed incoherent dictionary pair learning (InDPL) method in classification of a novel open-source database of Chinese numbers. Benchmarking results confirm that the InDPL algorithm offers enhanced classification accuracy, especially when the number of training samples is limited.

引用

页码：472 / 476

页数：5

共 35 条

[1] Fast and incoherent dictionary learning algorithms with application to fMRI [J].