Visual place recognition from a 3D laser LiDAR is one of the most active research areas in robotics. Especially, learning and recognition of scene descriptors, such as scan context descriptors that map 3D point clouds to 2D point clouds, is one of the promising research directions. Although the scan-context descriptor has a sufficiently high recognition performance, it is still expensive image data and cannot be handled with low-capacity non-deep models. In this paper, we explore the task of compressing the scan context descriptor model while maintaining its recognition performance. To this end, the proposed approach slightly modifies the off-the-shelf classifier model of convolutional neural networks (CNN) from its basis, by replacing the SoftMax part with a support vector machine (SVM). Experiments with publicly available NCLT dataset validate the effectiveness of the proposed approach.

1. [1] C. Masone and B. Caputo, “A survey on deep visual place recognition,” IEEE Access, Vol.9, pp. 19516-19547, 2021. https://doi.org/10.1109/ACCESS.2021.3054937
2. [2] X. Zhang, L. Wang, and Y. Su, “Visual place recognition: A survey from deep learning perspective,” Pattern Recognition, Vol.113, Article No.107760, 2021. https://doi.org/10.1016/j.patcog.2020.107760
3. [3] Z. Hong, Y. Petillot, D. Lane, Y. Miao, and S. Wang, “TextPlace: Visual place recognition and topological localization through reading scene texts,” 2019 IEEE/CVF Int. Conf. on Computer Vision (ICCV), pp. 2861-2870, 2019. https://doi.org/10.1109/ICCV.2019.00295
4. [4] M. Zaffar, S. Ehsan, M. Milford, and K. D. McDonald-Maier, “Memorable maps: A framework for re-defining places in visual place recognition,” IEEE Trans. on Intelligent Transportation Systems, Vol.22, No.12, pp. 7355-7369, 2021. https://doi.org/10.1109/TITS.2020.3001228
5. [5] G. Cha, S.-H. Sim, S. Park, and T. Oh, “LiDAR-based bridge displacement estimation using 3D spatial optimization,” Sensors, Vol.20, No.24, Article No.7117, 2020. https://doi.org/10.3390/s20247117
6. [6] J. Y. Chang, G. Moon, and K. M. Lee, “V2V-PoseNet: Voxel-to-voxel prediction network for accurate 3D hand and human pose estimation from a single depth map,” 2018 IEEE/CVF Conf. on Computer Vision and Pattern Recognition, pp. 5079-5088, 2018. https://doi.org/10.1109/CVPR.2018.00533
7. [7] A. P. Song, X. Y. Di, X. K. Xu, and Z. H. Song, “MeshGraphNet: An effective 3D polygon mesh recognition with topology reconstruction,” IEEE Access, Vol.8, pp. 205181-205189, 2020. https://doi.org/10.1109/ACCESS.2020.3037236
8. [8] J. Xu et al., “RPVNet: A deep and efficient range-point-voxel fusion network for LiDAR point cloud segmentation,” 2021 IEEE/CVF Int. Conf. on Computer Vision (ICCV), pp. 16024-16033, 2021. https://doi.org/10.1109/ICCV48922.2021.01572
9. [9] G. Chen, B. Bai, Z. Mao, and J. Dai, “Real-time road object segmentation using improved light-weight convolutional neural network based on 3D LiDAR point cloud,” Int. J. of Ad Hoc and Ubiquitous Computing, Vol.39, No.3, pp. 113-121, 2022. https://doi.org/10.1504/ijahuc.2022.121116
10. [10] L. Hui, M. Cheng, J. Xie, J. Yang, and M.-M. Cheng, “Efficient 3D point cloud feature learning for large-scale place recognition,” IEEE Trans. on Image Processing, Vol.31, pp. 1258-1270, 2022. https://doi.org/10.1109/TIP.2021.3136714
11. [11] P. Shi, Y. Zhang, and J. Li, “LiDAR-based place recognition for autonomous driving: A survey,” arXiv:2306.10561, 2023. https://doi.org/10.48550/arXiv.2306.10561
12. [12] K. Zhang, M. Hao, J. Wang, C. W. de Silva, and C. Fu, “Linked dynamic graph CNN: Learning on point cloud via linking hierarchical features,” arXiv:1904.10014, 2019. https://doi.org/10.48550/arXiv.1904.10014
13. [13] Y. Kojima, K. Tanaka, N. Yang, Y. Hirota, and K. Yamaguchi, “From comparison to retrieval: Scalable change retrieval from discriminatively learned deep three-dimensional neural codes,” 2019 IEEE Intelligent Transportation Systems Conf. (ITSC), pp. 789-795, 2019. https://doi.org/10.1109/ITSC.2019.8917170
14. [14] G. Kim and A. Kim, “Scan context: Egocentric spatial descriptor for place recognition within 3D point cloud map,” 2018 IEEE/RSJ Int. Conf. on Intelligent Robots and Systems (IROS), pp. 4802-4809, 2018. https://doi.org/10.1109/IROS.2018.8593953
15. [15] G. Kim, B. Park, and A. Kim, “1-day learning, 1-year localization: Long-term LiDAR localization using scan context image,” IEEE Robotics and Automation Letters, Vol.4, No.2, pp. 1948-1955, 2019. https://doi.org/10.1109/LRA.2019.2897340
16. [16] S. Y. Chaganti, I. Nanda, K. R. Pandi, T. G. N. R. S. N. Prudhvith, and N. Kumar, “Image classification using SVM and CNN,” 2020 Int. Conf. on Computer Science, Engineering and Applications (ICCSEA), 2020. https://doi.org/10.1109/ICCSEA49143.2020.9132851
17. [17] J. Komorowski, “MinkLoc3D: Point cloud based large-scale place recognition,” 2021 IEEE Winter Conf. on Applications of Computer Vision (WACV), pp. 1789-1798, 2021. https://doi.org/10.1109/WACV48630.2021.00183
18. [18] L. Hui, M. Cheng, J. Xie, J. Yang, and M.-M. Cheng, “Efficient 3D point cloud feature learning for large-scale place recognition,” IEEE Trans. on Image Processing, Vol.31, pp. 1258-1270, 2022. https://doi.org/10.1109/TIP.2021.3136714
19. [19] L. Hui, H. Yang, M. Cheng, J. Xie, and J. Yang, “Pyramid point cloud transformer for large-scale place recognition,” 2021 IEEE/CVF Int. Conf. on Computer Vision (ICCV), pp. 6078-6087, 2021. https://doi.org/10.1109/ICCV48922.2021.00604
20. [20] X. Kong et al., “Semantic graph based place recognition for 3D point clouds,” 2020 IEEE/RSJ Int. Conf. on Intelligent Robots and Systems (IROS), pp. 8216-8223, 2020. https://doi.org/10.1109/IROS45743.2020.9341060
21. [21] X. Xu et al., “DiSCO: Differentiable scan context with orientation,” IEEE Robotics and Automation Letters, Vol.6, No.2, pp. 2791-2798, 2021. https://doi.org/10.1109/LRA.2021.3060741
22. [22] L. Luo, S.-Y. Cao, B. Han, H.-L. Shen, and J. Li, “BVMatch: LiDAR-based place recognition using bird’s-eye view images,” IEEE Robotics and Automation Letters, Vol.6, No.3, pp. 6076-6083, 2021. https://doi.org/10.1109/LRA.2021.3091386
23. [23] L. Hoang, S.-H. Lee, and K.-R. Kwon, “A 3D shape recognition method using hybrid deep learning network CNN–SVM,” Electronics, Vol.9, No.4, Article No.649, 2020. https://doi.org/10.3390/electronics9040649
24. [24] H. Wang and X. Zhang, “Real-time vehicle detection and tracking using 3D LiDAR,” Asian J. of Control, Vol.24, No.3, pp. 1459-1469, 2022. https://doi.org/10.1002/asjc.2519
25. [25] F. Foroughi, Z. Chen, and J. Wang, “A CNN-based system for mobile robot navigation in indoor environments via visual localization with a small dataset,” World Electric Vehicle J., Vol.12, No.3, Article No.134, 2021. https://doi.org/10.3390/wevj12030134
26. [26] Y. Lecun, L. Bottou, Y. Bengio, and P. Haffner, “Gradient-based learning applied to document recognition,” Proc. of the IEEE, Vol.86, No.11, pp. 2278-2324, 1998. https://doi.org/10.1109/5.726791
27. [27] Z. Chen et al., “CAP-RAM: A charge-domain in-memory computing 6T-SRAM for accurate and precision-programmable CNN inference,” IEEE J. of Solid-State Circuits, Vol.56, No.6, pp. 1924-1935, 2021. https://doi.org/10.1109/JSSC.2021.3056447