Papers
Topics
Authors
Recent
Gemini 2.5 Flash
Gemini 2.5 Flash
162 tokens/sec
GPT-4o
7 tokens/sec
Gemini 2.5 Pro Pro
45 tokens/sec
o3 Pro
4 tokens/sec
GPT-4.1 Pro
38 tokens/sec
DeepSeek R1 via Azure Pro
28 tokens/sec
2000 character limit reached

Camera Agnostic Two-Head Network for Ego-Lane Inference (2404.12770v1)

Published 19 Apr 2024 in cs.CV, cs.LG, and cs.RO

Abstract: Vision-based ego-lane inference using High-Definition (HD) maps is essential in autonomous driving and advanced driver assistance systems. The traditional approach necessitates well-calibrated cameras, which confines variation of camera configuration, as the algorithm relies on intrinsic and extrinsic calibration. In this paper, we propose a learning-based ego-lane inference by directly estimating the ego-lane index from a single image. To enhance robust performance, our model incorporates the two-head structure inferring ego-lane in two perspectives simultaneously. Furthermore, we utilize an attention mechanism guided by vanishing point-and-line to adapt to changes in viewpoint without requiring accurate calibration. The high adaptability of our model was validated in diverse environments, devices, and camera mounting points and orientations.

Definition Search Book Streamline Icon: https://streamlinehq.com
References (25)
  1. P. Sarlin, C. Cadena, R. Siegwart, and M. Dymczyk, “From coarse to fine: Robust hierarchical localization at large scale,” in Proc. IEEE Conf. on Comput. Vision and Pattern Recog., 2019, pp. 12 716–12 725.
  2. P. Sarlin, D. DeTone, T. Malisiewicz, and A. Rabinovich, “Superglue: Learning feature matching with graph neural networks,” in Proc. IEEE Conf. on Comput. Vision and Pattern Recog., 2020, pp. 4937–4946.
  3. S. Yoon and A. Kim, “Line as a visual sentence: Context-aware line descriptor for visual localization,” IEEE Robot. and Automat. Lett., vol. 6, no. 4, pp. 8726–8733, 2021.
  4. S. Gao, J. Wan, Y. Ping, X. Zhang, S. Dong, Y. Yang, H. Ning, J. Li, and Y. Guo, “Pose refinement with joint optimization of visual points and lines,” in Proc. IEEE/RSJ Intl. Conf. on Intell. Robots and Sys., 2022, pp. 2888–2894.
  5. R. Asghar, M. Garzón, J. Lussereau, and C. Laugier, “Vehicle localization based on visual lane marking and topological map matching,” in Proc. IEEE Intl. Conf. on Robot. and Automat., 2020, pp. 258–264.
  6. M. He and R. R. Rajkumar, “Lanematch: A practical real-time localization method via lane-matching,” IEEE Robot. and Automat. Lett., vol. 7, no. 2, pp. 4408–4415, 2022.
  7. H. Wang, C. Xue, Y. Zhou, F. Wen, and H. Zhang, “Visual semantic localization based on HD map for autonomous vehicles in urban scenarios,” in Proc. IEEE Intl. Conf. on Robot. and Automat., 2021, pp. 11 255–11 261.
  8. S. Lee, S.-W. Kim, and S.-W. Seo, “Accurate ego-lane recognition utilizing multiple road characteristics in a bayesian network framework,” in Proc. IEEE Intell. Vehicle Symposium, 2015, pp. 543–548.
  9. A. L. Ballardini, D. Cattaneo, R. Izquierdo, I. Parra, M. Sotelo, and D. G. Sorrenti, “Ego-lane estimation by modeling lanes and sensor failures,” in Proc. IEEE Intell. Transport. Sys. Conf., 2017, pp. 1–7.
  10. A. Kasmi, D. Denis, R. Aufrere, and R. Chapuis, “Probabilistic framework for ego-lane determination,” in Proc. IEEE Intell. Vehicle Symposium, 2019, pp. 1746–1752.
  11. R. Arandjelovic, P. Gronát, A. Torii, T. Pajdla, and J. Sivic, “Netvlad: CNN architecture for weakly supervised place recognition,” IEEE Trans. Pattern Analysis and Machine Intell., vol. 40, no. 6, pp. 1437–1451, 2018.
  12. D. DeTone, T. Malisiewicz, and A. Rabinovich, “Superpoint: Self-supervised interest point detection and description,” in Proc. IEEE Conf. on Comput. Vision and Pattern Recog., 2018, pp. 224–236.
  13. M. J. Tyszkiewicz, P. Fua, and E. Trulls, “DISK: learning local features with policy gradient,” in Advances in Neural Information Processing Sys. Conf., 2020.
  14. K. Li, L. Wang, L. Liu, Q. Ran, K. Xu, and Y. Guo, “Decoupling makes weakly supervised local feature better,” in Proc. IEEE Conf. on Comput. Vision and Pattern Recog., 2022, pp. 15 817–15 827.
  15. X. Du and K. K. Tan, “Comprehensive and practical vision system for self-driving vehicle lane-level localization,” IEEE Trans. Image Processing, vol. 25, no. 5, pp. 2075–2088, 2016.
  16. M. Schreiber, C. Knöppel, and U. Franke, “Laneloc: Lane marking based localization using highly accurate maps,” in Proc. IEEE Intell. Vehicle Symposium, 2013, pp. 449–454.
  17. J. Jeong, Y. Cho, and A. Kim, “Hdmi-loc: Exploiting high definition map image for precise localization via bitwise particle filter,” IEEE Robot. and Automat. Lett., vol. 5, no. 4, pp. 6310–6317, 2020.
  18. T. Gao and H. K. Aghajan, “Self lane assignment using egocentric smart mobile camera for intelligent GPS navigation,” in Proc. IEEE Conf. on Comput. Vision and Pattern Recog., 2009, pp. 57–62.
  19. J. Rabe, M. Hübner, M. Necker, and C. Stiller, “Ego-lane estimation for downtown lane-level navigation,” in Proc. IEEE Intell. Vehicle Symposium, 2017, pp. 1152–1157.
  20. P. Bender, J. Ziegler, and C. Stiller, “Lanelets: Efficient map representation for autonomous driving,” in Proc. IEEE Intell. Vehicle Symposium, 2014, pp. 420–425.
  21. A. Vaswani, N. Shazeer, N. Parmar, J. Uszkoreit, L. Jones, A. N. Gomez, Ł. Kaiser, and I. Polosukhin, “Attention is all you need,” in Advances in Neural Information Processing Sys. Conf., 2017.
  22. Y. Zhou, H. Qi, J. Huang, and Y. Ma, “Neurvps: Neural vanishing point scanning via conic convolution,” Advances in Neural Information Processing Sys. Conf., vol. 32, 2019.
  23. M. Sensoy, L. Kaplan, and M. Kandemir, “Evidential deep learning to quantify classification uncertainty,” in Advances in Neural Information Processing Sys. Conf., vol. 31, 2018.
  24. K. He, X. Zhang, S. Ren, and J. Sun, “Deep residual learning for image recognition,” in Proc. IEEE Conf. on Comput. Vision and Pattern Recog., 2016, pp. 770–778.
  25. I. Loshchilov and F. Hutter, “Decoupled weight decay regularization,” arXiv preprint arXiv:1711.05101, 2017.

Summary

We haven't generated a summary for this paper yet.

Ai Sparkles Streamline Icon: https://streamlinehq.com Summarize Now