As an emerging visual task, vehicle re-identification refers to the identification of the same vehicle across multiple cameras. Herein, we propose a novel vehicle re-identification method that uses an improved ResNet-50 architecture and utilizes the topology information of a surveillance network to rerank the final results. In the training stage, we apply several data augmentation approaches to expand our training data and increase their diversity in a cost-effective manner. We reform the original RestNet-50 architecture by adding non-local blocks to implement the attention mechanism and replacing part of the batch normalization operations with instance batch normalization. After obtaining preliminary results from the proposed model, we use the reranking algorithm, whose core function is to improve the similarity scores of all images on the most likely path that the vehicle tends to appear to optimize the final results. Compared with most existing state-of-the-art methods, our method is lighter, requires less data annotation, and offers competitive performance.
[1] WOESLER R. Fast extraction of tra?c parame-ters and reidenti?cation of vehicles from video data [C]//Proceedings of the 2003 IEEE International Con-ference on Intelligent Transportation Systems. Shang-hai: IEEE, 2003: 774-778.
[2] SHAN Y, SAWHNEY H S, KUMAR R. Vehicle identi-?cation between non-overlapping cameras without di-rect feature matching [C]//Tenth IEEE International Conference on Computer Vision. Beijing: IEEE, 2005:
378- 385.
[3] ZHENG Q, LIANG C, FANG W H, et al. Car re-identi?cation from large scale images using semantic attributes [C]//2015 IEEE 17th International Work-shop on Multimedia Signal Processing. Xiamen: IEEE, 2015: 1-5.
[4] FERIS R S, SIDDIQUIE B, PETTERSON J, et al. Large-scale vehicle detection, indexing, and search in urban surveillance videos [J]. IEEE Transactions on Multimedia, 2012, 14(1): 28-42.
[5] ZAPLETAL D, HEROUT A. Vehicle re-identi?cation for automatic video tra?c surveillance [C]//2016 IEEE Conference on Computer Vision and Pattern Recognition Workshops. Las Vegas, NV: IEEE, 2016: 1568-1574.
[6] LIU X C, LIU W, MEI T, et al. A deep learning-based approach to progressive vehicle re-identi?cation for ur-ban surveillance[C]//European Conference on Com-puter Vision (ECCV 2016 ). Cham: Springer, 2016:
869- 884.
[7] LIU X C, LIU W, MA H D, et al. Large-scale vehicle re-identi?cation in urban surveillance videos [C]//2016 IEEE International Conference on Multimedia and Expo. Seattle, WA: IEEE, 2016: 1-6.
[8] LIU X C, LIU W, MEI T, et al. PROVID: Progressive and multimodal vehicle reidenti?cation for large-scale urban surveillance [J]. IEEE Transactions on Multime-dia, 2018, 20(3): 645-658.
[9] SZEGEDY C, LIU W, JIA Y Q, et al. Going deeper with convolutions [C]//2015 IEEE Conference on Computer Vision and Pattern Recognition.Boston, MA: IEEE, 2015: 1-9.
[10] WANG Z D, TANG L M, LIU X H, et al. Orienta-tion invariant feature embedding and spatial temporal regularization for vehicle Re-identi?cation [C]//2017 IEEE International Conference on Computer Vision. Venice: IEEE, 2017: 379-387.
[11] ZHOU Y, LIU L, SHAO L. Vehicle Re-identi?cation by deep hidden multi-view inference [J]. IEEE Trans-actions on Image Processing, 2018, 27(7): 3275-3287.
[12] ZHENG Z D, RUAN T, WEI Y C, et al. Ve-hicleNet: Learning robust visual representation for vehicle Re-identi?cation [EB/OL]. [2020-11-30]. https://arxiv.org/pdf/2004.06305.pdf. [13] ZHANG X Y, ZHANG R F, CAO J W, et al. Part-guided attention learning for vehi-cle instance retrieval [EB/OL]. [2020-11-30]. https://arxiv.org/pdf/1909.06023v4.pdf.
[14] KHORRAMSHAHI P, KUMAR A, PERI N, et al. A dual-path model with adaptive attention for vehicle Re-identi?cation [C]//2019 IEEE/CVF International Conference on Computer Vision. Seoul: IEEE, 2019: 6131-6140.
[15] CHEN T S, LIU C T, WU C W, et al. Orientation-aware vehicle re-identi?cation with semantics-guided part attention network [EB/OL]. [2020-11-30]. https://arxiv.org/pdf/2008.11423.pdf.
[16] HU J, SHEN L, SUN G. Squeeze-and-excitation net-works [C]//2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition. Salt Lake City, UT: IEEE, 2018: 7132-7141.
[17] SHEN Y T, XIAO T, LI H S, et al. Learning deep neural networks for vehicle Re-ID with visual-spatio-temporal path proposals [C]//2017 IEEE Interna-tional Conference on Computer Vision. Venice: IEEE, 2017: 1918-1927.
[18] ZHONG Z, ZHENG L, KANG G L, et al. Ran-dom erasing data augmentation [J]. Proceedings of the AAAI Conference on Arti?cial Intelligence, 2020, 34(7): 13001-13008.
[19] HE KM, ZHANGX Y,REN S Q,et al. Deep residual learning for image recognition [C]//2016 IEEE Con-ference on Computer Vision and Pattern Recognition. Las Vegas, NV: IEEE, 2016: 770-778.
[20] LUO H, GU Y Z, LIAO X Y, et al. Bag of tricks and a strong baseline for deep person re-identi?cation [C]//2019 IEEE/CVF Conference on Computer Vi-sion and Pattern Recognition Workshops.LongBeach, CA: IEEE, 2019: 1487-1495.
[21] HE L X, LIAO X Y, LIU W, et al. FastReID: A pytorch toolbox for real-world person re-identi?cation [EB/OL]. [2020-11-30]. https://arxiv.org/pdf/ 2006.02631.pdf.
[22] PANXG, LUOP,SHIJP,etal. Two atonce: Enhancing learning and generalization capacities via IBN-net[C]//Proceedings of the European Conference on Computer Vision (ECCV ). Munich: [s.n.], 2018:
464- 479.
[23] ULYANOV D, VEDALDI A, LEMPITSKY V. Instance normalization: The missing ingredi-ent for fast stylization [EB/OL]. [2020-11-30]. https://arxiv.org/pdf/1607.08022.pdf.
[24] IOFFE S, SZEGEDY C. Batch normalization: Accelerating deep network training by reducing internal covariate shift [EB/OL]. [2020-11-30]. https://arxiv.org/pdf/1502.03167.pdf.
[25] WOO S, PARK J, LEE J Y, et al. CBAM: Convolu-tional block attention module [M]//Computer vision
– ECCV 2018. Cham: Springer International Publish-ing, 2018: 3-19.
[26] WANG F, JIANG M Q, QIAN C, et al. Residual atten-tion network for image classi?cation [C]//2017 IEEEConference on Computer Vision and Pattern Recogni-tion. Honolulu, HI: IEEE, 2017: 6450-6458. [27] HAAS R, HOFFMANN M. Chordless paths through three vertices [J]. Theoretical Computer Science, 2006, 351(3): 360-371.
[28] LIU H Y, TIAN Y H, WANG Y W, et al. Deep relative distance learning: Tell the di?erence between similar vehicles [C]//2016 IEEE Conference on Computer Vi-sion and Pattern Recognition. Las Vegas, NV: IEEE, 2016: 2167-2175.
[29] GUO H Y, ZHAO C Y, LIU Z W, et al. Learning coarse-to-?ne structured feature embedding for vehi-cle re-identi?cation[C]//Thirty-Second AAAI Confer-ence on Arti?cial Intelligence. New Orleans, Louisiana: AAAI, 2018.
[30] LOU Y H, BAI Y, LIU J, et al. VERI-wild: A large dataset and a new method for vehicle re-identi?cation in the wild [C]//2019 IEEE/CVF Conference on Com-puter Vision and Pattern Recognition.LongBeach, CA: IEEE, 2019: 3230-3238.
[31] ZHONG Z, ZHENG L, CAO D L, et al. Re-ranking person re-identi?cation with k-reciprocal encoding [C]//2017 IEEE Conference on Computer Vision and Pattern Recognition. Honolulu, HI: IEEE, 2017: 3652-3661. [32] BAI S, BAI X. Sparse contextual activation for e?-cient visual re-ranking [J]. IEEE Transactions on Im-age Processing, 2016, 25(3): 1056-1069.
[33] SARFRAZ M S, SCHUMANN A, EBERLE A, et al. A pose-sensitive embedding for person re-identi?cation with expanded cross neighborhood Re-ranking [C]//2018 IEEE/CVF Conference on Com-puter Vision and Pattern Recognition. Salt Lake City, UT: IEEE, 2018: 420-429.
[34] GUO R P, LI C G, LI Y H, et al. Density-adaptive kernel based re-ranking for person re-identi?cation [C]//2018 24th International Conference on Pattern Recognition. Beijing: IEEE, 2018: 982-987.
[35] HE B, LI J, ZHAO Y F, et al. Part-regularized near-duplicate vehicle re-identi?cation [C]//2019 IEEE/CVF Conference on Computer Vision and Pattern Recognition. Long Beach, CA: IEEE, 2019: 3992-4000.
