Abstract
Achieving high quality recognition result for video and natural scene images that contain both standard 2D text as well as decorative 3D text is challenging. Methods developed for 2D text may fail for 3D text due to the presence of pixels representing shadow and depth in the 3D text. This work aims at classification of 2D and 3D texts in video or scene images such that one can choose an appropriate method in the classified text for achieving better results. The proposed method explores Generalized Gradient Vector Flow (GGVF) for finding dominant points for input 2D and 3D text line images based on opposite direction symmetry. For each dominant point, our approach finds distance between neighbor points and plots a histogram to choose points which contribute to the highest peak as candidates. Distance symmetry between a candidate point and its neighbor points is checked and if a candidate point is visited twice, a common point is created. Statistical features such as the mean and standard deviation of the common points and candidate points are extracted to feed to Neural Network (NN) for classification. Experimental results on dataset of 2D-3D text line images and the dataset collected from standard natural scene images show that the proposed method outperforms exiting methods. Furthermore, recognition experiments before and after classification show recognition performance improves significantly as a result of applying our method.
This is a preview of subscription content, log in via an institution to check access.
Access this chapter
Tax calculation will be finalised at checkout
Purchases are for personal use only
References
Raghunandan, K.S., Shivakumara, P., Roy, S., Kumar, G.H., Pal, U., Lu, T.: Multi-script-oriented text detection and recognition in video/scene/born digital images. IEEE Trans. CSVT 29, 1145–1162 (2019)
Roy, S., Shivakumara, P., Pal, U., Lu, T., Kumar, G.H.: Delaunay triangulation based text detection from multi-view images of natural scene. PRL 129, 92–100 (2020)
Shivakumara, P., Raghavendra, R., Qin, L., Raja, K.B., Lu, T., Pal, U.: A new multi-modal approach to bib number/text detection and recognition in Marathon images. PR 61, 479–491 (2017)
Xue, M., Shivakumara, P., Zhang, C., Lu, T., Pal, U.: Curved text detection in blurred/non-blurred video/scene images. MTAP 78(18), 25629–25653 (2019). https://doi.org/10.1007/s11042-019-7721-2
Tian, S., Yin, X.C., Su, Y., Hao, H.W.: A unified framework for tracking based text detection and recognition from web videos. IEEE Trans. PAMI 40, 542–554 (2018)
Shi, B., Yang, M., Wang, X., Luy, P., Yao, C., Bai, X.: ASTER: an attentional scene text recognizer with flexible rectification. IEEE Trans. PAMI 41, 2035–2048 (2019)
Luo, C., Jin, L., Sun, Z.: MORAN: a multi-object rectified attention network for scene text recognition. PR 90, 109–118 (2019)
Baek, Y., Lee, B., Han, D., Yun, S., Lee, H.: Character region awareness for text detection. In: Proceedings of CVPR (2019)
Cheng, Z., Xu, Y., Bai, F., Niu, Y.: AON: towards arbitrarily-oriented text recognition. In: Proceedings of CVPR, pp. 5571–5579 (2018)
Qi, X., Chen, Y., Xiao, R., Li, C.G., Zou, Q., Cui, S.: A novel joint character categorization and localization approach for character level scene text recognition. In: Proceedings of ICDARW, pp. 83–90 (2019)
Rong, X., Yi, C., Tian, Y.: Unambiguous scene text segmentation with referring expression comprehension. IEEE Trans. IP 29, 591–601 (2020)
Villamizar, M., Canevert, O., Odobez, J.M.: Multi-scale sequential network for semantic text segmentation and localization. PRL 129, 63–69 (2020)
Feng, W., He, W., Yin, F., Zhang, X.Y., Liu, C.L.: TextDragon: an end-to-end framework for arbitrary shaped text spotting. In: Proceedings of ICCV, pp. 9076–9085 (2019)
Xu, J., Shivakumara, P., Lu, T., Tan, C.L., Uchida, S.: A new method for multi-oriented graphics-scene-3D text classification in video. PR 49, 19–42 (2016)
Zhong, W., Raj, A.N.J., Shivakumara, P., Zhuang, Z., Lu, T., Pal, U.: A new shadow detection and depth removal method for 3d text recognition in scene images. In: Proceedings of ICIMT, pp. 277–281 (2018)
Khare, V., Shivakumara, P., Chan, C.S., Lu, T., Meng, L.K., Woon, H.H., Blumenstein, M.: A novel character segmentation-reconstruction approach for license plate recognition. ESWA 131, 219–239 (2019)
Zhu, S., Gao, R.: A novel generalized gradient vector flow snake model using minimal surface and component-normalized method for medical image segmentation. BSPC 26, 1–10 (2016)
Silva, I.N.D., Spatti, D.H., Flauzino, R.A., Liboni, L.H.B., Reis Alves, S.F.D.: Artificial Neural Networks, vol. 39. Springer, Heidelberg (2017). https://doi.org/10.1007/978-3-319-43162-8
Narayan, S.: The generalized sigmoid activation function: competitive supervised learning. IS 99, 69–82 (1997)
Nasr, G.E., Badr, E.A., Joun, C.: Cross entropy error function in neural networks: forecasting gasoline demand. In: Proceedings of FLAIRS, pp. 381–384 (2002)
Kingma, P.D., Bai, J.L.: Adam: a method for stochastic optimization. In: Proceedings of ICLR, pp. 1–15 (2015)
Keras: Deep learning library for theano and tensorflow (2015). https://keras.io/
Author information
Authors and Affiliations
Corresponding author
Editor information
Editors and Affiliations
Rights and permissions
Copyright information
© 2020 Springer Nature Switzerland AG
About this paper
Cite this paper
Nandanwar, L., Shivakumara, P., Kumar, A., Lu, T., Pal, U., Lopresti, D. (2020). A New Common Points Detection Method for Classification of 2D and 3D Texts in Video/Scene Images. In: Bai, X., Karatzas, D., Lopresti, D. (eds) Document Analysis Systems. DAS 2020. Lecture Notes in Computer Science(), vol 12116. Springer, Cham. https://doi.org/10.1007/978-3-030-57058-3_36
Download citation
DOI: https://doi.org/10.1007/978-3-030-57058-3_36
Published:
Publisher Name: Springer, Cham
Print ISBN: 978-3-030-57057-6
Online ISBN: 978-3-030-57058-3
eBook Packages: Computer ScienceComputer Science (R0)
