Effectively Leveraging Visual Context to Detect Texts in Natural Scenes

Detecting texts in natural scenes is challenging because of large variation in size and layout of texts and strong distractions from background clutters. Leveraging contextual information is crucial in boosting the detection accuracy. In this paper, we construct a conditional random field (CRF) to utilize visual context that helps enhance true detections and suppress false alarms. Unlike previous works, the pairwise potentials in our model encode three different compatibility/repulsion relationships among character candidates under two different layout scenarios, and the unary potentials are obtained from multi-class recognition confidence of individual character candidates. In addition, we use easy texts to help recover difficult ones in an iterative manner. Due to these efforts, our method outperforms state-of-the-art text detection algorithms on the challenging ICDAR dataset.

[1]  Simon M. Lucas,et al.  ICDAR 2003 robust reading competitions , 2003, Seventh International Conference on Document Analysis and Recognition, 2003. Proceedings..

[2]  Andrew McCallum,et al.  Conditional Random Fields: Probabilistic Models for Segmenting and Labeling Sequence Data , 2001, ICML.

[3]  Jiri Matas,et al.  A Method for Text Localization and Recognition in Real-World Images , 2010, ACCV.

[4]  Leo Breiman,et al.  Random Forests , 2001, Machine Learning.

[5]  H. Tran,et al.  A Novel Approach for Text Detection in Images Using Structural Features , 2005, ICAPR.

[6]  Xujun Peng,et al.  Text Extraction from Video Using Conditional Random Fields , 2011, 2011 International Conference on Document Analysis and Recognition.

[7]  Chunheng Wang,et al.  Text detection in images based on unsupervised classification of edge-based features , 2005, Eighth International Conference on Document Analysis and Recognition (ICDAR'05).

[8]  Kongqiao Wang,et al.  An Improved Scene Text Extraction Method Using Conditional Random Field and Optical Character Recognition , 2011, 2011 International Conference on Document Analysis and Recognition.

[9]  Sudeep Sarkar,et al.  Robust outdoor text detection using text intensity and shape features , 2008, 2008 19th International Conference on Pattern Recognition.

[10]  Bill Triggs,et al.  Histograms of oriented gradients for human detection , 2005, 2005 IEEE Computer Society Conference on Computer Vision and Pattern Recognition (CVPR'05).

[11]  Yonatan Wexler,et al.  Detecting text in natural scenes with stroke width transform , 2010, 2010 IEEE Computer Society Conference on Computer Vision and Pattern Recognition.

[12]  Chew Lim Tan,et al.  Ieee Transactions on Pattern Analysis and Machine Intelligence, Manuscript Id a Laplacian Approach to Multi-oriented Text Detection in Video , 2022 .

[13]  Alan L. Yuille,et al.  Detecting and reading text in natural scenes , 2004, Proceedings of the 2004 IEEE Computer Society Conference on Computer Vision and Pattern Recognition, 2004. CVPR 2004..

[14]  Jagath Samarabandu,et al.  Multiscale Edge-Based Text Extraction from Complex Images , 2006, 2006 IEEE International Conference on Multimedia and Expo.

[15]  Cheng-Lin Liu,et al.  Text Localization in Natural Scene Images Based on Conditional Random Field , 2009, 2009 10th International Conference on Document Analysis and Recognition.

[16]  T. Heskes.  Stable Fixed Points of Loopy Belief Propagation Are Minima of the Bethe Free Energy , 2002 .

[17]  Palaiahnakote Shivakumara,et al.  2009 10th International Conference on Document Analysis and Recognition A Gradient Difference based Technique for Video Text Detection , 2022 .

[18]  Chucai Yi,et al.  Text String Detection From Natural Scenes by Structure-Based Partition and Grouping , 2011, IEEE Transactions on Image Processing.

[19]  Ching Y. Suen,et al.  Text detection from scene images using sparse representation , 2008, 2008 19th International Conference on Pattern Recognition.

[20]  Lionel Prevost,et al.  2009 10th International Conference on Document Analysis and Recognition Text Detection and Localization in Complex Scene Images using Constrained AdaBoost Algorithm , 2022 .

[21]  Lionel Prevost,et al.  A cascade detector for text detection in natural scene images , 2008, 2008 19th International Conference on Pattern Recognition.