Abstract
In this paper, an effective scene text detection algorithm is proposed based on Maximally Stable External Region (MSER) and deep convolutional network. This algorithm is competitive with the best existing scheme widely used on benchmark International Conference on Document Analysis and Recognition (ICDAR) 2013. The main contribution of the paper lies in two aspects. Firstly, an efficient text/background classifier is designed on the basis of convolutional neural network, which outperforms the artificial design features. What’s more, synthesis data is used for constructing the training set so as to avoid over-fitting; secondly, an analysis method based on MSER is presented to drop out the majority background patches and extract valid character areas. On the basis of analysing connected components, the size of candidate character is determined instead of the time-consuming scale scanning.
Sign-in/Register to access full text options 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a callDisclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.
