TY  - JOUR
ID  - scholars13862
KW  - Character recognition; Network architecture
KW  -  Benchmark datasets; Deep neural network; Document recognition; Documents analysis; International conference on document analyse and recognition dataset; Localisation; Multi-oriented text; Spatial transformer network; Spatial transformer network-OCR; Text region
KW  -  Deep neural networks
N2  - Developing systems for interpreting visuals, such as images, videos is really challenging but important task to be developed and applied on benchmark datasets. This study solves the very challenge by using STN-OCR model consisting of deep neural networks (DNN) and Spatial Transformer Networks (STNs). The network architecture of this study consists of two stages: localization network and recognition network. In the localization network it finds and localizes text regions and generates sampling grid. Whereas, in the recognition network, text regions will be input and then this network learns to recognize text including low resolution, curved and multi-oriented text. Deep learning-based approaches require a lot of data for training effectively, therefore, this study has used two benchmark datasets, Street View House Numbers (SVHN) and International Conference on Document Analysis and Recognition (ICDAR) 2015 to evaluate the system. The STN-OCR model achieves better results than literature on these datasets. © 2020 Science and Information Organization.
IS  - 4
VL  - 11
JF  - International Journal of Advanced Computer Science and Applications
A1  - Katper, S.H.
A1  - Gilal, A.R.
A1  - Waqas, A.
A1  - Alshanqiti, A.
A1  - Alsughayyir, A.
A1  - Jaafar, J.
UR  - https://www.scopus.com/inward/record.uri?eid=2-s2.0-85085521236&doi=10.14569%2fIJACSA.2020.0110424&partnerID=40&md5=20229642338dccd786a3cf530a4f0869
Y1  - 2020///
SP  - 178
TI  - Deep neural networks nombined with STN for multi-oriented text detection and recognition
N1  - cited By 9
AV  - none
EP  - 184
PB  - Science and Information Organization
SN  - 2158107X
ER  -