自动驾驶场景类间相似特征自适应分类网络

doi:10.3778/j.issn.1673-9418.2403033

摘要/Abstract

摘要： 解决类间相似度问题是自动驾驶场景分类研究中一项充满挑战的任务，主要研究在相似度较高的真实复杂交通场景中，利用网络学习目标特征的差异性，并构建特征之间整体关联性进行场景分类。提出一种多尺度自适应特征筛选的自动驾驶场景分类网络。采用双重多尺度特征提取模块预处理，初步提取不同尺度下的类间相似特征；设计了特征分化筛选模块完成场景相似特征筛选，使网络更关注不同场景类别的典型易区分特征；将特征筛选结果和多尺度特征图共同传递至特征融合分类模块进行场景分类，捕捉场景特征之间的关联性；由自适应学习算法通过输出结果动态调整训练参数，加快网络收敛速度并提升精度。所提方法在三种数据集BDD100k、BDD100k+和自制数据集上与现有网络方法进行比较，相较Top2网络在精度上分别领先了3.29%、5.59%、12.65%（相对），实验结果表明了所提方法的有效性，并展现了很好的泛化能力。提出的场景分类方法旨在学习不同复杂场景类别下的典型易区分的特征及其关联性，降低多目标类间相似的影响，使得在真实交通场景数据集中场景分类结果更加准确。

关键词: 自动驾驶, 场景分类, 类间相似, 多尺度结构, 特征筛选, 自适应训练

Abstract: Addressing the issue of inter-class similarity is a challenging task in the research of autonomous driving scene classification, which primarily focuses on learning the distinctive features of targets in real-world complex traffic scenarios with high similarity, and constructing the overall correlation between features for scene classification. To this end, a multi-scale adaptive feature selection network for autonomous driving scene classification is proposed. Initially, a dual multi-scale feature extraction module is utilized for preliminary processing to extract inter-class similar features at different scales. Subsequently, a feature differentiation screening module is designed to complete the screening of scene-similar features, enabling the network to focus more on the typical and easily distinguishable features of different scene categories. Then, the feature screening results and multi-scale feature maps are transferred to the feature fusion classification module for scene classification, and the correlation between scene features is captured. Finally, an adaptive learning algorithm dynamically adjusts the training parameters through the output results, accelerating the network's convergence speed and improving accuracy. The proposed method is compared with existing network methods on three datasets: BDD100k, BDD100k+ and self-made dataset. Compared with the Top2 networks, it leads in accuracy by 3.29%, 5.59% and 12.65% (relatively), respectively. Experimental results demonstrate the effectiveness of the proposed method and its strong generalization capability. The scene classification method presented in this paper aims to learn the typical and easily distinguishable features and their correlations under different complex scene categories, reducing the impact of inter-class similarity among multiple targets, thereby making the scene classification results in real-world traffic scenario datasets more accurate.

Key words: autonomous driving, scene classification, inter-class similarity, multi-scale structure, feature screening, adaptive training

姜彦吉, 冯宇宙, 董浩, 田佳琳. 自动驾驶场景类间相似特征自适应分类网络[J]. 计算机科学与探索, 2024, 18(11): 3051-3064.

JIANG Yanji, FENG Yuzhou, DONG Hao, TIAN Jialin. Adaptive Classification Network for Similar Features Between Classes in Automatic Driving Scenarios[J]. Journal of Frontiers of Computer Science and Technology, 2024, 18(11): 3051-3064.

参考文献

[1] PENG T, SU L, ZHANG R, et al. A new safe lane change trajectory model and collision avoidance control method for automatic driving vehicles[J]. Expert Systems with Applications, 2020, 141: 112953.
[2] YAMADA M, UEDA K, HORIBA I, et al. Discrimination of the road condition toward understanding of vehicle driving environments[J]. IEEE Transactions on Intelligent Transportation Systems, 2021, 2(1): 26-31.
[3] TONG W, CHEN W, HAN W, et al. Channel-attention based DenseNet network for remote sensing image scene classification[J]. IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing, 2020, 13: 4121-4132.
[4] JING C, HUANG L, CAI S, et al. Interclass similarity transfer for imbalanced aerial scene classification[J]. IEEE Geoscience and Remote Sensing Letters, 2023, 20: 1-5.
[5] LATTE M V, SHIDNAL S, ANAMI B, et al. A combined color and texture features based methodology for recognition of crop field image[J]. International Journal of Signal Processing, Image Processing and Pattern Recognition, 2015, 8(2): 287-302.
[6] CHEN L C, PAPANDREOU G, KOKKINOS I, et al. Semantic image segmentation with deep convolutional nets and fully connected CRFs[EB/OL]. [2024-01-21]. https://arxiv.org/abs/1412.7062.
[7] ZHAO W, DU S. Scene classification using multi-scale deeply described visual words[J]. International Journal of Remote Sensing, 2016, 37(17): 4119-4131.
[8] LIU Y, ZHONG Y, QIN Q. Scene classification based on multiscale convolutional neural network[J]. IEEE Transactions on Geoscience and Remote Sensing, 2018, 56(12): 7109-7121.
[9] 张睿, 杨义鑫, 李阳, 等. 自监督学习下小样本遥感图像场景分类[J]. 中国图象图形学报, 2022, 27(11): 3371-3381.
ZHANG R, YANG Y X, LI Y, et al. Self-supervised learning based few-shot remote sensing scene image classification[J]. Journal of Image and Graphics, 2022, 27(11): 3371-3381.
[10] 邬满, 文莉莉, 孙苗. 注意力机制海洋场景图像理解算法[J]. 计算机工程与应用, 2022, 58(10): 231-239.
WU M, WEN L L, SUN M. Attention mechanism image understanding algorithm of ocean scene[J]. Computer Engineering and Applications, 2022, 58(10): 231-239.
[11] 王光, 陶燕, 沈慧芳, 等. 基于多特征融合与CELM的场景分类算法[J]. 计算机工程与应用, 2022, 58(1): 232-240.
WANG G, TAO Y, SHEN H F, et al. Multi-feature fusion and constrained extreme learning machine for scene classification[J]. Computer Engineering and Applications, 2022, 58(1): 232-240.
[12] 杜鹏, 张有明, 朱郑州, 等. 迁移学习在低资源场景实体识别中的应用研究[J]. 计算机科学与探索, 2023, 17(4): 912-921.
DU P, ZHANG Y M, ZHU Z Z, et al. Study on application of transfer learning in entity recognition of low resource environment[J]. Journal of Frontiers of Computer Science and Technology, 2023, 17(4): 912-921.
[13] 张多纳, 赵宏佳, 鲁远耀, 等. 融入注意力机制的小样本遥感图像场景分类[J]. 计算机工程与应用, 2024, 60(4): 173-182.
ZHANG D N, ZHAO H J, LU Y Y, et al. Few-shot scene classification with attention mechanism in remote sensing[J]. Computer Engineering and Applications, 2024, 60(4): 173-182.
[14] CHEN L, ZHAN W, TIAN W, et al. Deep integration: a multi-label architecture for road scene recognition[J]. IEEE Transactions on Image Processing, 2019, 28(10): 4883-4898.
[15] ZHENG K, NAJI H A H. Road scene segmentation based on deep learning[J]. IEEE Access, 2020, 8: 140964-140971.
[16] CHEN W, OUYANG S, TONG W, et al. GCSANet: a global context spatial attention deep learning network for remote sensing scene classification[J]. IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing, 2022, 15: 1150-1162.
[17] SAFFARI M, KHODAYAR M, JALALI S M J. Sparse adversarial unsupervised domain adaptation with deep dictionary learning for traffic scene classification[J]. IEEE Transactions on Emerging Topics in Computational Intelligence, 2023, 7(4): 1139-1150.
[18] YU F, CHEN H, WANG X, et al. BDD100k: a diverse driving dataset for heterogeneous multitask learning[C]//Proceedings of the 2020 IEEE/CVF Conference on Computer Vision and Pattern Recognition. Piscataway: IEEE, 2020: 2636-2645.
[19] SHEN C, ZHAO X, FAN X, et al. Multi-receptive field graph convolutional neural networks for pedestrian detection[J]. IET Intelligent Transport Systems, 2019, 13(9): 1319-1328.
[20] XU H, SRIVASTAVA G. Automatic recognition algorithm of traffic signs based on convolution neural network[J]. Multimedia Tools and Applications, 2020, 79(17/18): 11551-11565.
[21] DALAL N, TRIGGS B. Histograms of oriented gradients for human detection[C]//Proceedings of the 2005 IEEE Computer Society Conference on Computer Vision and Pattern Recognition. Washington: IEEE Computer Society, 2005: 886-893.
[22] WOO S, PARK J, LEE J Y, et al. CBAM: convolutional block attention module[C]//Proceedings of the 2018 European Conference on Computer Vision. Cham: Springer, 2018: 3-19.
[23] YU F, KOLTUN V, FUNKHOUSER T. Dilated residual networks[C]//Proceedings of the 2017 IEEE Conference on Computer Vision and Pattern Recognition. Washington: IEEE Computer Society, 2017: 472-480.
[24] NAMBURU A, SELVARAJ P, MOHAN S, et al. Forest fire identification in UAV imagery using X-MobileNet[J]. Electronics, 2023, 12(3): 733.
[25] CHOLLET F. Xception: deep learning with depthwise separable convolutions[C]//Proceedings of the 2017 IEEE Conference on Computer Vision and Pattern Recognition. Washington: IEEE Computer Society, 2017: 1251-1258.
[26] CHANG H S, LEARNED-MILLER E, MCCALLUM A. Active bias: training more accurate neural networks by emphasizing high variance samples[C]//Advances in Neural Information Processing Systems 30, Long Beach, Dec 4-9, 2017: 1002-1012.
[27] 张德, 李国璋, 王怀光, 等. 位姿估计自适应学习率的改进[J]. 电子测量与仪器学报, 2019, 33(6): 51-58.
ZHANG D, LI G Z, WANG H G, et al. Improved adaptive learning rate of pose estimation[J]. Journal of Electronic Measurement and Instrumentation, 2019, 33(6): 51-58.
[28] WANG W, HU Y, ZOU T, et al. A new image classification approach via improved MobileNet models with local receptive field expansion in shallow layers[J]. Computational Intelligence and Neuroscience, 2020. DOI: 10.1155/2020/8817849.
[29] LIU S, TIAN G, XU Y. A novel scene classification model combining ResNet based transfer learning and data augmentation with a filter[J]. Neurocomputing, 2019, 338: 191-206.
[30] HUANG X, YANG R, WANG Q, et al. A novel method for real-time ATR system of AUV based on Attention-MobileNet-V3 network and pixel correction algorithm[J]. Ocean Engineering, 2023, 270: 113403.
[31] PENG C, LIU Y, YUAN X, et al. Research of image recognition method based on enhanced inception-ResNet-V2[J]. Multimedia Tools and Applications, 2022, 81(24): 34345-34365.
[32] ALHICHRI H, ALSWAYED A S, BAZI Y, et al. Classification of remote sensing images using EfficientNet-B3 CNN model with attention[J]. IEEE Access, 2021, 9: 14078-14094.
[33] RATEKE T, JUSTEN K A, VON WANGENHEIM A. Road surface classification with images captured from low cost camera-road traversing knowledge (RTK) dataset[J]. Revista de Informática Teórica e Aplicada, 2019, 26(3): 50-64.
[34] MYLAVARAPU S, SANDHU M, VIJAYAN P, et al. Understanding dynamic scenes using graph convolution networks[C]//Proceedings of the 2020 IEEE/RSJ International Conference on Intelligent Robots and Systems. Piscataway: IEEE, 2020: 8279-8286.