TY - GEN
T1 - Scene character recognition via bag-of-words model
T2 - 5th International Conference on Communications, Signal Processing and Systems, CSPS 2016
AU - Zhang, Zhong
AU - Wang, Hong
AU - Liu, Shuang
N1 - Publisher Copyright:
© Springer Nature Singapore Pte Ltd. 2018.
PY - 2018
Y1 - 2018
N2 - In this paper, we focus on the feature representation methods under the framework of bag-of-words model (BoW) for scene character recognition. We investigate three kinds of methods: the original BoW methods, the stroke-based methods, and the context-based methods. Specifically, various feature representation methods are introduced, their merits and demerits are explored, and existing problems are discussed. Finally, we empirically evaluate them on several widely used databases (Chars74k, ICDAR2003 and SVHN).
AB - In this paper, we focus on the feature representation methods under the framework of bag-of-words model (BoW) for scene character recognition. We investigate three kinds of methods: the original BoW methods, the stroke-based methods, and the context-based methods. Specifically, various feature representation methods are introduced, their merits and demerits are explored, and existing problems are discussed. Finally, we empirically evaluate them on several widely used databases (Chars74k, ICDAR2003 and SVHN).
KW - Bag of words
KW - Feature representation
KW - Scene character recognition
UR - http://www.scopus.com/inward/record.url?scp=85034260465&partnerID=8YFLogxK
U2 - 10.1007/978-981-10-3229-5_88
DO - 10.1007/978-981-10-3229-5_88
M3 - Conference contribution
AN - SCOPUS:85034260465
SN - 9789811032288
T3 - Lecture Notes in Electrical Engineering
SP - 819
EP - 826
BT - Communications, Signal Processing, and Systems - Proceedings of the 2016 International Conference on Communications, Signal Processing, and Systems
A2 - Mu, Jiasong
A2 - Wang, Wei
A2 - Zhang, Baoju
A2 - Liang, Qilian
PB - Springer Verlag
Y2 - 21 October 2016 through 22 October 2016
ER -