Scene content area and division are two indispensable and testing research issues in the field of PC vision. This paper proposes a novel strategy for scene content revelation and division in light of fell convolution neural Networks (CNNs). In this system, a CNN based substance careful cheerful substance district (CTR) extraction show (named recognizable proof orchestrate, DNet) is arranged and arranged using both the edges and the whole territories of substance, with which coarse CTRs are recognized. A CNN based CTR refinement show (named division organize, SNet) is then created to section the coarse CTRs into substance to get the refined CTRs. With DNet and SNet, numerous less CTRs are removed than with regular philosophies while all the more bona fide content areas are kept. The refined CTRs are finally requested using a CNN based CTR game plan illustrate (named gathering framework, CNet) to get the last substance locale. This paper proposes a novel scene content area procedure by using distinctive convolution neural frameworks. This technique contains three phases including content careful CTR extraction, CTR refinement, and CTR course of action.

Scene Text detection, scene text segmentation, text-aware candidate text region extraction, candidate text region refinement, candidate text region classification

[1] A. Shahab, F. Shafait, and A. Dengel, “ICDAR 2011 robust reading competition challenge 2: Reading text in scene images,” in Proc. Int. Conf. Document Anal. Recognit., Sep. 2011, pp. 1491–1496.
[2] D. Karatzas et al., “ICDAR 2013 robust reading competition,” in Proc. Int. Conf. Document Anal. Recognit., Aug. 2013, pp. 1484–1493.
[3] D. Karatzas et al., “ICDAR 2015 competition on robust reading,” in Proc. Int. Conf. Document Anal. Recognit., Aug. 2015, pp. 1156–1160.
[4] B. Epshtein, E. Ofek, and Y. Wexler, “Detecting text in natural scenes with stroke width transform,” in Proc. IEEE Conf. Comput. Vis. Pattern Recognit., Jun. 2010, pp. 2963–2970.
[5] J. Matas, O. Chum, M. Urban, and T. Pajdla, “Robust wide-baseline stereo from maximally stable extremal regions,” Image Vis. Comput., vol. 22, no. 10, pp. 761–767, 2004.
[6] C. Yi and Y. Tian, “Text string detection from natural scenes by structure-based partition and grouping,” IEEE Trans. Image Process., vol. 20, no. 9, pp. 2594–2605, Sep. 2011.
[7] C. Yao, X. Bai, W. Liu, Y. Ma, and Z. Tu, “Detecting texts of arbitrary orientations in natural images,” in Proc. IEEE Conf. Comput. Vis. Pattern Recognit., Jun. 2012, pp. 1083–1090.
[8] L. Neumann and J. Matas, “Real-time scene text localization and recognition,” in Proc. IEEE Conf. Comput. Vis. Pattern Recognit., Jun. 2012, pp. 3538–3545.
[9] W. Huang, Z. Lin, J. Yang, and J. Wang, “Text localization in natural images using stroke feature transform and text covariance descriptors,” in Proc. Int. Conf. Comput. Vis., Dec. 2013, pp. 1241–1248.
[10] L. Neumann and J. Matas, “Scene text localization and recognition with oriented stroke detection,” in Proc. Int. Conf. Comput. Vis., Dec. 2013, pp. 97–104.
[11] C. Shi, C. Wang, B. Xiao, Y. Zhang, and S. Gao, “Scene text detection using graph model built upon maximally stable extremal regions,” Pattern Recognit. Lett., vol. 34, no. 2, pp. 107–116, Jan. 2013.
[12] X.-C. Yin, X. Yin, K. Huang, and H.-W. Hao, “Robust text detection in natural scene images,” IEEE Trans. Pattern Anal. Mach. Intell., vol. 36, no. 5, pp. 970–983, May 2014.
[13] L. Neumann and J. Matas, “A method for text localization and recognition
in real-world images,” in Proc. Asian Conf. Comput. Vis., 2010, pp. 770–783.
[14] W. Huang, Y. Qiao, and X. Tang, “Robust scene text detection with convolution neural network induced MSER trees,” in Proc. Eur. Conf. Comput. Vis., 2014, pp. 497–511.
[15] Y. Li, W. Jia, C. Shen, and A. van den Hengel, “Characterness: An indicator of text in the wild,” IEEE Trans. Image Process., vol. 23, no. 4, pp. 1666–1677, Apr. 2014.