一种基于对抗学习的实时跟踪模型设计

doi:10.19678/j.issn.1000-3428.0057920

摘要/Abstract

摘要： 目标跟踪指在视频帧中找到感兴趣目标的运动位置，广泛应用于环境感知、安防监控和无人驾驶等领域。为进行高效的目标跟踪，建立一种基于对抗学习和特征压缩的相关滤波器目标跟踪模型。为了同时兼顾精度与速度，在模型中引入特征提取优化、特征压缩和特征聚合等步骤。在提取图像特征前，采用对抗学习方法解决特征提取模型中训练数据与任务数据分布不匹配的问题。在特征压缩阶段，应用双通道自编码器结构和特征聚合来增强模型对图像风格的泛化能力。实验结果表明，与非实时跟踪算法相比，该模型在精度损失不超过3%的情况下能取得明显的速度提升，其跟踪速度高达103FPS。

关键词: 目标跟踪, 对抗学习, 自编码器, 相关滤波器, 表示学习

Abstract: Target tracking is the process of finding the position of the moving target in the video frame, and is widely used in the fields of environmental perception, security monitoring and unmanned driving.To improve the efficiency of target tracking, this paper proposes a target tracking model using correlation filter based on adversarial learning and feature compression.The model includes the optimization of feature extraction, feature compression and feature aggregation, which improve both the accuracy and the speed of the model.Before feature extraction, adversarial learning is used to solve the problem of the mismatch of training data and task data distribution in the feature extraction model.In the stage of feature compression, a two-way autoencoder structure and feature aggregation are used to enhance the generalization of the image style.Experimental results show that compared with the non-real-time tracking algorithm, the model significantly improves the speed, with the tracking rate reaching 103FPS.At the same time, the loss of the model accuracy is within 3%.

Key words: target tracking, adversarial learning, autoencoder, correlation filter, representation learning

中图分类号:

TP391

李志鹏, 张睿. 一种基于对抗学习的实时跟踪模型设计[J]. 计算机工程, 2021, 47(6): 262-270.

LI Zhipeng, ZHANG Rui. Design of a Real-Time Tracking Model Based on Adversarial Learning[J]. Computer Engineering, 2021, 47(6): 262-270.

https://www.ecice06.com/CN/Y2021/V47/I6/262

图/表 9

20210618184744

20210618184748

20210618184751

20210618184754

20210618184757

20210618184801

20210618184804

20210618184808

20210618184811

参考文献

[1] NAM H,HAN B.Learning multi-domain convolutional neural networks for visual tracking[C]//Proceedings of IEEE Conference on Computer Vision and Pattern Recognition.Washington D.C.,USA:IEEE Press,2016:4293-4302.
[2] TAO R,GAVVES E,SMEULDERS A W M.Siamese instance search for tracking[C]//Proceedings of IEEE Conference on Computer Vision and Pattern Recognition.Washington D.C.,USA:IEEE Press,2016:1420-1429.
[3] DANELLJAN M,HAGER G,SHAHBAZ KHAN F,et al.Convolutional features for correlation filter based visual tracking[C]//Proceedings of IEEE International Conference on Computer Vision Workshops.Washington D.C.,USA:IEEE Press,2015:58-66.
[4] DANELLJAN M,BHAT G,SHAHBAZ KHAN F,et al.ECO:efficient convolution operators for tracking[C]//Proceedings of IEEE Conference on Computer Vision and Pattern Recognition.Washington D.C.,USA:IEEE Press,2017:6638-6646.
[5] WANG L,OUYANG W,WANG X,et al.STCT:sequentially training convolutional networks for visual tracking[C]//Proceedings of IEEE Conference on Computer Vision and Pattern Recognition.Washington D.C.,USA:IEEE Press,2016:1373-1381.
[6] CHATFIELD K,SIMONYAN K,VEDALDI A,et al.Return of the devil in the details:delving deep into convolutional nets[EB/OL].[2020-02-05].http://de.arxiv.org/pdf/1405.3531.
[7] SAITO K,WATANABE K,USHIKU Y,et al.Maximum classifier discrepancy for unsupervised domain adaptation[C]//Proceedings of IEEE Conference on Computer Vision and Pattern Recognition.Washington D.C.,USA:IEEE Press,2018:3723-3732.
[8] YAN X,YANG J,SOHN K,et al.Attribute2Image:conditional image generation from visual attributes[C]//Proceedings of European Conference on Computer Vision.Berlin,Germany:Springer,2016:776-791.
[9] DU B,XIONG W,WU J,et al.Stacked convolutional denoising auto-encoders for feature representation[J].IEEE Transactions on Cybernetics,2016,47(4):1017-1027.
[10] RUSSAKOVSKY O,DENG J,SU H,et al.Imagenet large scale visual recognition challenge[J].International Journal of Computer Vision,2015,115(3):211-252.
[11] HENRIQUES J F,CASEIRO R,MARTINS P,et al.High-speed tracking with kernelized correlation filters[J].IEEE Transactions on Pattern Analysis and Machine Intelligence,2014,37(3):583-596.
[12] LI Y,ZHU J.A scale adaptive kernel correlation filter tracker with feature integration[C]//Proceedings of European Conference on Computer Vision.Berlin,Germany:Springer,2014:254-265.
[13] DANELLJAN M,HÄGER G,KHAN F,et al.Accurate scale estimation for robust visual tracking[C]//Proceedings of British Machine Vision Conference.Washington D.C.,USA:IEEE Press,2014:15-22.
[14] ZHANG K,ZHANG L,LIU Q,et al.Fast visual tracking via dense spatio-temporal context learning[C]//Proceedings of European Conference on Computer Vision.Berlin,Germany:Springer,2014:127-141.
[15] MA C,HUANG J B,YANG X,et al.Hierarchical convolutional features for visual tracking[C]//Proceedings of IEEE International Conference on Computer Vision.Washington D.C.,USA:IEEE Press,2015:3074-3082.
[16] DANELLJAN M,HAGER G,SHAHBAZ KHAN F,et al.Learning spatially regularized correlation filters for visual tracking[C]//Proceedings of IEEE International Conference on Computer Vision.Washington D.C.,USA:IEEE Press,2015:4310-4318.
[17] WANG N,LI S,GUPTA A,et al.Transferring rich feature hierarchies for robust visual tracking[EB/OL].[2020-02-05].https://arxiv.org/pdf/1501.04587.pdf.
[18] NAM H,HAN B.Learning multi-domain convolutional neural networks for visual tracking[C]//Proceedings of IEEE Conference on Computer Vision and Pattern Recognition.Washington D.C.,USA:IEEE Press,2016:4293-4302.
[19] WANG L,OUYANG W,WANG X,et al.Visual tracking with fully convolutional networks[C]//Proceedings of IEEE International Conference on Computer Vision.Washington D.C.,USA:IEEE Press,2015:3119-3127.
[20] FAN H,LING H.SANet:structure-aware network for visual tracking[C]//Proceedings of IEEE Conference on Computer Vision and Pattern Recognition Workshops.Washington D.C.,USA:IEEE Press,2017:42-49.
[21] YUN S,CHOI J,YOO Y,et al.Action-decision networks for visual tracking with deep reinforcement learning[C]//Proceedings of IEEE Conference on Computer Vision and Pattern Recognition.Washington D.C.,USA:IEEE Press,2017:2711-2720.
[22] BEN-DAVID S,BLITZER J,CRAMMER K,et al.A theory of learning from different domains[J].Machine Learning,2010,79(1/2):151-175.
[23] GOODFELLOW I,POUGET-ABADIE J,MIRZA M,et al.Generative adversarial nets[EB/OL].[2020-02-05].https://papers.nips.cc/paper/2014/file/5ca3e9b122f61f8f06494c97b1afccf3-Paper.pdf.
[24] KINGMA D P,BA J.Adam:a method for stochastic optimization[EB/OL].[2020-02-05].http://de.arxiv.org/pdf/1412.6980.
[25] WU Y,LIM J,YANG M H.Object tracking benchmark[J].IEEE Transactions on Pattern Analysis and Machine Intelligence,2015,37(9):1834-1848.
[26] CHOI J,JIN CHANG H,JEONG J,et al.Visual tracking using attention-modulated disintegration and integration[C]//Proceedings of IEEE Conference on Computer Vision and Pattern Recognition.Washington D.C.,USA:IEEE Press,2016:4321-4330.
[27] BERTINETTO L,VALMADRE J,HENRIQUES J F,et al.Fully-convolutional siamese networks for object tracking[C]//Proceedings of European Conference on Computer Vision.Berlin,Germany:Springer,2016:850-865.
[28] MAATEN L,HINTON G.Visualizing data using t-SNE[J].Journal of Machine Learning Research,2008,9:2579-2605.

选择文件类型/文献管理软件名称

选择包含的内容