Chinese Journal of Mechanical Engineering >

2021 , Vol. 34 >Issue 5: 78 - 78

DOI: https://doi.org/10.1186/s10033-021-00598-9

Original Article

ML-ANet: A Transfer Learning Approach Using Adaptation Network for Multi-label Image Classification in Autonomous Driving

  • Guofa Li ,
  • Zefeng Ji ,
  • Yunlong Chang ,
  • Shen Li ,
  • Xingda Qu ,
  • Dongpu Cao
Expand
  • 1. Institute of Human Factors and Ergonomics, College of Mechatronics and Control Engineering, Shenzhen University, Shenzhen, 518060, China;
    2. Department of Mechanical and Mechatronics Engineering, University of Waterloo, Waterloo, ON, N2L 3G1, Canada;
    3. Interactive Entertainment Group, Tencent, Shanghai, 200233, China;
    4. Traffic Operations and Safety Laboratory, University of Wisconsin-Madison, Madison, WI, 53715, USA

Received date: 2020-06-01

  Revised date: 2021-02-22

  Online published: 2022-03-22

Supported by

Supported by Shenzhen Fundamental Research Fund of China (Grant No. JCYJ20190808142613246), National Natural Science Foundation of China (Grant No. 51805332), and Young Elite Scientists Sponsorship Program funded by the China Society of Automotive Engineers.

Fold

Abstract

To reduce the discrepancy between the source and target domains, a new multi-label adaptation network (ML-ANet) based on multiple kernel variants with maximum mean discrepancies is proposed in this paper. The hidden representations of the task-specific layers in ML-ANet are embedded in the reproducing kernel Hilbert space (RKHS) so that the mean-embeddings of specific features in different domains could be precisely matched. Multiple kernel functions are used to improve feature distribution efficiency for explicit mean embedding matching, which can further reduce domain discrepancy. Adverse weather and cross-camera adaptation examinations are conducted to verify the effectiveness of our proposed ML-ANet. The results show that our proposed ML-ANet achieves higher accuracies than the compared state-of-the-art methods for multi-label image classification in both the adverse weather adaptation and cross-camera adaptation experiments. These results indicate that ML-ANet can alleviate the reliance on fully labeled training data and improve the accuracy of multi-label image classification in various domain shift scenarios.

Key words： Autonomous vehicles; Deep learning; Image classification; Multi-label learning; Transfer learning

Cite this article

Guofa Li , Zefeng Ji , Yunlong Chang , Shen Li , Xingda Qu , Dongpu Cao . ML-ANet: A Transfer Learning Approach Using Adaptation Network for Multi-label Image Classification in Autonomous Driving[J]. Chinese Journal of Mechanical Engineering, 2021 , 34(5) : 78 -78 . DOI: 10.1186/s10033-021-00598-9

References

[1] H Gao, B Cheng, J Wang, et al. Object classification using CNN-based fusion of vision and LIDAR in autonomous vehicle environment. IEEE Transactions on Industrial Informatics, 2018, 16(9):4224-4231.
[2] G Li, Y Yang, X Qu, et al. A deep learning based image enhancement approach for autonomous driving at night. Knowledge-Based Systems, 2021, 213:106617.
[3] G Li, Y Yang, T Zhang, et al. Risk assessment based collision avoidance decision-making for autonomous vehicles in multi-scenarios. Transportation Research Part C:Emerging Technologies, 2021, 122:102820.
[4] G Li, S Li, S Li, et al. Deep reinforcement learning enabled decision-making for autonomous driving at intersections. Automotive Innovation, 2020, 3:374-385.
[5] Y Chen, W Li, C Sakaridis, et al. Domain adaptive faster R-CNN for object detection in the wild. Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, 2018:3339-3348.
[6] G Li, S E Li, R Zou, et al. Detection of road traffic participants using cost-effective arrayed ultrasonic sensors in low-speed traffic situations. Mechanical Systems and Signal Processing, 2019, 132:535-545.
[7] S E Li, G Li, J Yu, et al. Kalman filter-based tracking of moving objects using linear ultrasonic sensor array for road vehicles. Mechanical Systems and Signal Processing, 2018, 98:173-189.
[8] X Zhang, Z Chen, Q M J Wu, et al. Fast semantic segmentation for scene perception. IEEE Transactions on Industrial Informatics, 2018, 15(2):1183-1192.
[9] L Mou, P Ghamisi, X X Zhu. Deep recurrent neural networks for hyperspectral image classification. IEEE Transactions on Geoscience and Remote Sensing, 2017, 55(7):3639-3655.
[10] Z Yan, W Liu, S. Wen, et al. Multi-label image classification by feature attention network. IEEE Access, 2019, 7:98005-98013.
[11] M L Zhang, Z H Zhou, ML-KNN:A lazy learning approach to multi-label learning. Pattern Recognition, 2007, 40(7):2038-2048.
[12] A Elisseeff, J Weston. A kernel method for multi-labelled classification. Neural Information Processing Systems, 2001, 14:681-687.
[13] M L Zhang, Z H Zhou. A review on multi-label learning algorithms. IEEE Transactions on Knowledge and Data Engineering, 2013, 26(8):1819-1837.
[14] Y Wei, W Xia, M Lin, et al. HCP:A flexible CNN framework for multi-label image classification. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2015, 38(9):1901-1907.
[15] J Wang, Y Yang, J Mao, et al. CNN-RNN:A unified framework for multi-label image classification. Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, 2016:2285-2294.
[16] J Zhang, Q Wu, C Shen, et al. Multilabel image classification with regional latent semantic dependencies. IEEE Transactions on Multimedia, 2018, 20(10):2801-2813.
[17] L Song, J Liu, B Qian, et al. A deep multi-modal CNN for multi-instance multi-label image classification. IEEE Transactions on Image Processing, 2018, 27(12):6025-6038.
[18] F C Heilbron, J C Niebles. Collecting and annotating human activities in web videos. Proceedings of International Conference on Multimedia Retrieval, 2014:377-384.
[19] G Li, Y Chen, D Cao, et al. Extraction of descriptive driving patterns from driving data using unsupervised algorithms. Mechanical Systems and Signal Processing, 2021, 156:107589.
[20] S J Pan, Q Yang. A survey on transfer learning. IEEE Transactions on Knowledge and Data Engineering, 2009, 22(10):1345-1359.
[21] S J Pan, I W Tsang, J T Kwok, et al. Domain adaptation via transfer component analysis. IEEE Transactions on Neural Networks, 2010, 22(2):199-210.
[22] M Long, J Wang, G Ding, et al. Transfer feature learning with joint distribution adaptation. Proceedings of the IEEE International Conference on Computer Vision, 2013:2200-2207.
[23] J Wang, Y Chen, S Hao, et al. Balanced distribution adaptation for transfer learning. IEEE International Conference on Data Mining, 2017:1129-1134.
[24] J Wang, Y Chen, H Yu, et al. Easy transfer learning by exploiting intra-domain structures. IEEE International Conference on Multimedia and Expo, 2019:1210-1215.
[25] J Yosinski, J Clune, Y Bengio, et al. How transferable are features in deep neural networks?. arXiv preprint arXiv, 2014:arxiv:1411.1792.
[26] N Tajbakhsh, J Y Shin, S R Gurudu, et al. Convolutional neural networks for medical image analysis:Full training or fine tuning? IEEE Transactions on Medical Imaging, 2016, 35(5):1299-1312.
[27] X Zhang, F X Yu, S F Chang, et al. Deep transfer network:Unsupervised domain adaptation. arXiv preprint arXiv, 2015:arxiv:1503.00591.
[28] E Tzeng, J Hoffman, N Zhang, et al. Deep domain confusion:Maximizing for domain invariance. arXiv preprint arXiv, 2014:arxiv:1412.3474.
[29] A Gretton, K M Borgwardt, M J Rasch, et al. A kernel two-sample test. The Journal of Machine Learning Research, 2012, 13(1):723-773.
[30] C Sakaridis, D Dai, L Van Gool. Semantic foggy scene understanding with synthetic data. International Journal of Computer Vision, 2018, 126(9):973-992.
[31] G Li, Y Wang, F Zhu, et al. Drivers' visual scanning behavior at signalized and unsignalized intersections:A naturalistic driving study in China. Journal of Safety Research, 2019, 71:219-229.
[32] G Li, W Lai, X Sui, et al. Influence of traffic congestion on driver behavior in post-congestion driving. Accident Analysis and Prevention, 2020, 141:105508.
[33] G Li, S E Li, B Cheng, et al. Estimation of driving style in naturalistic highway traffic using maneuver transition probabilities. Transportation Research Part C:Emerging Technologies, 2017, 74:113-125.
[34] M Cordts, M Omran, S Ramos, et al. The cityscapes dataset for semantic urban scene understanding. Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, 2016:3213-3223.
[35] A Geiger, P Lenz, R Urtasun. Are we ready for autonomous driving? the kitti vision benchmark suite. IEEE Conference on Computer Vision and Pattern Recognition, 2012:3354-3361.
[36] G Li, Y Yang, X Qu. Deep learning approaches on pedestrian detection in hazy weather. IEEE Transactions on Industrial Electronics, 2019, 67(10):8889-8899.
[37] Q Wen, Z Luo, R Chen, et al. Deep learning approaches on defect detection in high resolution aerial images of insulators. Sensors, 2021, 21:1033.
[38] Z Wojna, V Ferrari, S Guadarrama, et al. The devil is in the decoder. British Machine Vision Conference, 2017:1-13.
[39] M Long, Y Cao, J Wang, et al. Learning transferable features with deep adaptation networks. International Conference on Machine Learning, 2015:97-105.
[40] G Li, Y Lin, X Qu. An infrared and visible image fusion method based on multi-scale transformation and norm optimization. Information Fusion, 2021:https://doi.org/10.1016/j.inffus.2021.02.008
[41] G Li, H Xie, X Qu, et al. Detection of road objects with mall appearance in images for autonomous driving in various traffic situations using a deep learning based approach. IEEE Access, 2020, 8:211146-211172.
[42] G Li, Y Liao, Q Guo, et al. Traffic crash characteristics in Shenzhen, China from 2014 to 2016. International Journal of Environmental Research and Public Health, 2021, 18:1176.
Options
Outlines

/

〈 〉

　 Copyright © Chinese Journal of Mechanical Engineering, All Rights Reserved.
Powered by Beijing Magtech Co. Ltd,.