# A Combinatorial Solution to Point Symbol Recognition

^{1}

^{2}

^{*}

## Abstract

**:**

## 1. Introduction

## 2. Related Works

#### 2.1. Point Symbols in the Topographic Map

#### 2.2. Application of Deep Learning in Object Recognition

## 3. Method

Algorithm Framework Procedure of Point Symbol Recognition |

Step 1 Prescreening the topographic maps. |

Step 1.1 The topographic map is processed to extract the black sub-layouts. |

Step 1.2 Based on the method of judging the minimum bounding rectangle, the connected region of suspected point symbols should be obtained. |

Step 2 Training the model. |

Step 2.1 Pretraining the AlexNet networks based on MNIST database by transfer learning to obtain the pretraining model. |

Step 2.2 Pretaining the pretraining model based on the point-symbols data to obtain the final model. |

Step 3 Recognition of the point-symbols. |

The test images of point symbols input the final model to test the recognition accuracy. |

#### 3.1. The Prescreen Based on the Characteristics of Point Symbols

#### 3.2. The Point Symbol Recognition Based on AlexNet

## 4. Experiment Analysis and Results

#### 4.1. The Point Symbols Dataset

#### 4.2. The Comparison of Algorithms

#### 4.2.1. The Prescreening of Point Symbols

#### 4.2.2. The Recognition of Point Symbols

#### 4.3. Runtime Analysis

## 5. Conclusions

## Author Contributions

## Funding

## Acknowledgments

## Conflicts of Interest

## References

- Lladós, J.; Valveny, E.; Sánchez, G.; Martí, E. Symbol recognition: Current advances and perspectives. Lect. Notes Comput. Sci.
**2001**, 2390, 104–127. [Google Scholar] - Camassa, R.; Kuang, D.; Long, L. A geodesic landmark shooting algorithm for template matching and its applications. SIAM J. Imag. Sci.
**2017**, 10, 303–334. [Google Scholar] [CrossRef] - Wang, Q.; Zhang, F.; Li, X. Optimal clustering framework for hyperspectral band selection. IEEE Trans. Geosci. Remote Sens.
**2018**, 99, 1–13. [Google Scholar] [CrossRef] - Miao, Q.; Xu, P.; Liu, T.; Yang, Y.; Zhang, J.; Li, W. Linear feature separation from topographic maps using energy density and the shear transform. IEEE Trans. Image Process.
**2013**, 22, 1546–1556. [Google Scholar] - Rahman, S.Z.A.; Abdullah, S.N.H.S.; Nazri, M.Z.B.A. The analysis for gait energy image based on statistical methods. In Proceedings of the International Conference on Advances in Electrical, Electronic and Systems Engineering, Putrajaya, Malaysia, 14–16 November 2016; pp. 125–128. [Google Scholar]
- Sadahiro, Y. A statistical method for determining the size of map labels. Theor. Appl. Gis
**2009**, 3, 33–44. [Google Scholar] [CrossRef] - Woods, M.L.; Leon, M.; Perezcarrasco, R.; Barnes, C.P. A statistical approach reveals designs for the most robust stochastic gene oscillators. ACS Synth. Biol.
**2016**, 5, 459–470. [Google Scholar] [CrossRef] [PubMed] - Coles, R. Race and Family: A Structural Approach; Rowman & Littlefield: Lanham, MD, USA, 2006. [Google Scholar]
- Cai, J.; Liu, Z.Q. Integration of structural and statistical information for unconstrained handwritten numeral recognition. IEEE Trans. Pattern Anal. Mach. Intell.
**1999**, 21, 263–270. [Google Scholar] - Wang, Q.; He, X.; Li, X. Locality and structure regularized low rank representation for hyperspectral image classification. IEEE Trans. Geosci. Remote Sens.
**2018**, 99, 1–13. [Google Scholar] [CrossRef] - Osadchy, M.; Cun, Y.L.; Miller, M.L. Synergistic face detection and pose estimation with energy-based models. J. Mach. Learn. Res.
**2006**, 8, 1197–1215. [Google Scholar] - Ullah, I.; Lee, H.J. An approach of locating korean vehicle license plate based on mathematical morphology and geometrical features. In Proceedings of the 2016 International Conference on Computational Science and Computational Intelligence (CSCI), Las Vegas, NV, USA, 15–17 December 2016; pp. 836–840. [Google Scholar]
- Miao, Q.; Liu, T.; Song, J.; Gong, M.; Yang, Y. Guided superpixel method for topographic map processing. IEEE Trans. Geosci. Remote Sens.
**2016**, 54, 6265–6279. [Google Scholar] [CrossRef] - Wang, N. Recognition based on BP neural network map symbol. Digit. Technol. Appl.
**2014**, 11, 50–52. [Google Scholar] - Chen, F.; Hu, P.; Wang, X.; Fan, Q. Automatic recognition of point symbols on topographic maps. Geoinform
**2007**, 6751, 675113. [Google Scholar] - Wang, Q.; Yuan, Z.; Du, Q.; Li, X. GETNET: A general end-to-end 2-D CNN framework for hyperspectral image change detection. IEEE Trans. Geosci. Remote Sens.
**2018**, 99, 1–11. [Google Scholar] [CrossRef] - Zhang, T.; Yin, Y.L.; Liu, X.L. Automatic recognition of point-shaped symbols in scanned maps. Eng. Surv. Mapp.
**2000**, 3, 15–18. [Google Scholar] - Szegedy, C.; Liu, W.; Jia, Y.; Sermanet, P.; Reed, S.; Anguelov, D.; Erhan, D.; Vanhoucke, V.; Rabinovich, A. Going deeper with convolutions. In Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, Boston, MA, USA, 7–12 June 2015; pp. 1–9. [Google Scholar]
- Jisheng, B.I.; Tian, L.; Zhang, G. Intelligent recognition of map point symbols based on cognitive theory. Hydrogr. Surv. Chart.
**2016**, 1, 65–67. [Google Scholar] - Nass, A.; Gasselt, S.V. Dynamic Cartography: A Concept for Multidimensional Point Symbols. In Progress in Cartography; Gartner, G., Jobst, M., Huang, H., Eds.; Springer International Publishing: Cham, Switzerland, 2016; pp. 17–30. [Google Scholar]
- Krizhevsky, A.; Sutskever, I.; Hinton, G.E. ImageNet classification with deep convolutional neural networks. In Proceedings of the International Conference on Neural Information Processing Systems, Lake Tahoe, NV, USA, 12–14 November 2012. [Google Scholar]
- Nakada, M.; Wang, H.; Terzopoulos, D. AcFR: Active face recognition using convolutional neural networks. In Proceedings of the Computer Vision and Pattern Recognition Workshops, Honolulu, HI, USA, 21–26 July 2017; pp. 35–40. [Google Scholar]
- Liu, Y.; Huang, H.; Cao, J.; Huang, T. Convolutional neural networks-based intelligent recognition of Chinese license plates. Soft Comput.
**2017**, 22, 1–17. [Google Scholar] [CrossRef] - Ibañez, R.; Soria, A.; Teyseyre, A.R.; Berdun, L.; Campo, M.R. A Comparative Study of Machine Learning Techniques for Gesture Recognition Using Kinect. In Handbook of Research on HumanComputer Interfaces, Developments, and Applications; Rodrigues, J., Cardoso, P., Monteiro, J., Figueiredo, M., Eds.; IGI Global: Philadelphia, PA, USA, 2016; pp. 1–22. [Google Scholar]
- Miao, Q.; Xu, P.; Li, X.; Song, J.; Li, W.; Yang, Y. The recognition of the point symbols in the scanned topographic maps. IEEE Trans. Image Process.
**2017**, 26, 2751–2766. [Google Scholar] [CrossRef] [PubMed] - Simonyan, K.; Zisserman, A. Very deep convolutional networks for large-scale image recognition. arXiv, 2014; arXiv:1409.1556. [Google Scholar]

**Figure 2.**The figure shows the comparison of map changes after color segmentation, as follows: (

**a**) the original image of the topography map; (

**b**) the sub-layout images of the topography map.

**Figure 3.**This figure plays the comparison of map changes after prescreening. They are listed as (

**a**) description of figure after binarization; (

**b**) description of the figure that extracts the black sub-layouts and eliminates the noise.

**Figure 9.**The prescreening of point symbols is presented in the figure. They are listed as (

**a**) the original map; (

**b**) the grid pattern; (

**c**) the connected region pattern with prescreening; (

**d**) the connected region pattern.

Category | 1 | 2 | 3 | 4 | 5 | 6 | 7 | 8 | 9 |
---|---|---|---|---|---|---|---|---|---|

Symbols | |||||||||

Number | 250 | 205 | 200 | 238 | 240 | 243 | 204 | 219 | 212 |

Category | 1 | 2 | 3 | 4 | 5 | 6 | 7 | 8 | 9 |
---|---|---|---|---|---|---|---|---|---|

Ration | 1 | 1 | 0.935 | 0.963 | 0.902 | 0.932 | 0.984 | 0.896 | 0.82 |

LeNet (Pre-Trained) | VGG-16 (Pre-Trained) | AlexNet (Not Pre-Trained) | AlexNet (Pre-Trained) | BP Networks | SLS-GH | |
---|---|---|---|---|---|---|

The 1st test image | 93.5 | 90.61 | 96.76 | 98.85 | 89.65 | 98.57 |

The 2nd test image | 90.33 | 95.32 | 96.76 | 98.97 | 83.68 | 98.28 |

The 3rd test image | 92.87 | 94.35 | 95.68 | 99.56 | 89.45 | 98.56 |

The 4th test image | 92.56 | 96.61 | 94.84 | 98.49 | 81.23 | 97.65 |

The 5th test image | 91.32 | 93.89 | 95.15 | 98.96 | 83.12 | 97.76 |

Average | 92.12 | 94.45 | 95.84 | 98.97 | 85.43 | 98.16 |

AlexNet—Not Pre-Trained (ms) | BP Networks (ms) | SLS-GH (ms) | |
---|---|---|---|

The 1st test image | 256 | 564 | 890 |

The 2nd> test image | 196 | 456 | 501 |

The 3rd test image | 396 | 1064 | 1460 |

The 4th test image | 231 | 256 | 328 |

The 5th test image | 485 | 536 | 689 |

© 2018 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).

## Share and Cite

**MDPI and ACS Style**

Quan, Y.; Shi, Y.; Miao, Q.; Qi, Y.
A Combinatorial Solution to Point Symbol Recognition. *Sensors* **2018**, *18*, 3403.
https://doi.org/10.3390/s18103403

**AMA Style**

Quan Y, Shi Y, Miao Q, Qi Y.
A Combinatorial Solution to Point Symbol Recognition. *Sensors*. 2018; 18(10):3403.
https://doi.org/10.3390/s18103403

**Chicago/Turabian Style**

Quan, Yining, Yuanyuan Shi, Qiguang Miao, and Yutao Qi.
2018. "A Combinatorial Solution to Point Symbol Recognition" *Sensors* 18, no. 10: 3403.
https://doi.org/10.3390/s18103403