Multi-lingual Scene Text Detection by Local Histogram Analysis and Selection of Optimal Area for MSER

The problem of scene text detection has been quite intriguing for the research fraternity due to wide scope of applications. In recent times, some robust technique like Maximally Stable Extremal Region (MSER) has been developed and have gained immense popularity in detection and localization of text components in the wild. However, the problem of an optimal threshold and range of area in MSER has been addressed by very few researchers. In this work, we address this problem by dynamic thresholding through local histogram analysis and formulating a generalized range of the area for the MSER function. The detected MSERs are then analyzed by observing their region properties to distinguish text components from non-text ones and finally bounding the text components. The proposed technique is evaluated on a set of 200 images with multi-lingual text comprising English, Bangla, Hindi and Oriya languages.

[1]  Huizhong Chen,et al.  Robust text detection in natural images with edge-enhanced Maximally Stable Extremal Regions , 2011, 2011 18th IEEE International Conference on Image Processing.

[2]  Palaiahnakote Shivakumara,et al.  Script independent approach for multi-oriented text detection in scene image , 2017, Neurocomputing.

[3]  Jun Zhang,et al.  Multi-Orientation Scene Text Detection with Adaptive Clustering , 2015, IEEE Transactions on Pattern Analysis and Machine Intelligence.

[4]  Jiřı́ Matas,et al.  Real-time scene text localization and recognition , 2012, 2012 IEEE Conference on Computer Vision and Pattern Recognition.

[5]  Dimosthenis Karatzas,et al.  Multi-script Text Extraction from Natural Scenes , 2013, 2013 12th International Conference on Document Analysis and Recognition.

[6]  Dimosthenis Karatzas,et al.  TextProposals: A text-specific selective search algorithm for word spotting in the wild , 2016, Pattern Recognit..

[7]  Partha Pratim Roy,et al.  Script independent scene text segmentation using fast stroke width transform and GrabCut , 2015, 2015 3rd IAPR Asian Conference on Pattern Recognition (ACPR).

[8]  Deepak Kumar,et al.  Multi-script robust reading competition in ICDAR 2013 , 2013, MOCR '13.

[9]  Luis Miguel Bergasa,et al.  Text location in complex images , 2012, Proceedings of the 21st International Conference on Pattern Recognition (ICPR2012).

[10]  Subhadip Basu,et al.  Multi-lingual Text Localization from Camera Captured Images Based on Foreground Homogenity Analysis , 2018, Advances in Intelligent Systems and Computing.

[11]  Huchuan Lu,et al.  Scene text detection via stroke width , 2012, Proceedings of the 21st International Conference on Pattern Recognition (ICPR2012).

[12]  Yao Li,et al.  Characterness: An Indicator of Text in the Wild , 2013, IEEE Transactions on Image Processing.

[13]  Jiri Matas,et al.  Robust wide-baseline stereo from maximally stable extremal regions , 2004, Image Vis. Comput..

[14]  Chunna Tian,et al.  Natural scene text detection with MC-MR candidate extraction and coarse-to-fine filtering , 2017, Neurocomputing.

[15]  Nabendu Chaki,et al.  A new image binarization method using iterative partitioning , 2012, Machine Vision and Applications.