CMNet: A Connect-and-Merge Convolutional Neural Network for Fast Vehicle Detection in Urban Traffic Surveillance
Open accessAbstract
A wide variety of vehicle detection approaches using the deep convolutional neural network (CNN) have achieved great success in recent years. However, the existing CNN-based feature extraction algorithms, especially residual network, cannot obtain powerful semantic information in the vehicle detection task, and thus suffer from the problem of a missing detection, error detection, or repeated detection. In this paper, we present a connect-and-merge convolutional neural network (CMNet) for fast detecting vehicles in complex scenes. First, we propose a connect-and-merge residual network (CMRN) for performing feature extraction. Specifically, the CMRN assembles residual branches in parallel through a connect-and-merge mapping: Connect the input to the outputs of two residual branches separately (Connect), and merge the outputs of the connection as the input of the subsequent residual block (Merge), respectively. Second, we present a multi-scale prediction network (MSPN) to accurately regress the vehicle shape and classify vehicle fine-grained categories. In addition, the feature maps from the CMRN are merged with their corresponding upsampled features from the MSPN using concatenation. Thus, we can improve the information flow of the framework and make it easy to train. The experimental results on the KITTI dataset and the UA-DETRAC dataset demonstrate that the CMNet can achieve efficient detection performance in terms of accuracy and speed for the real-world traffic surveillance data.
Integrity signals
Registry
Cites
Statistics
Claims the paper makes
Claims other papers make about this one
Each is checked against this paper’s own text: does it say what the citing paper says it does?
Our works achieve better performance compared with recent robust detectors [31] and [32].
“Our works achieve better performance compared with recent robust detectors, such as [31] and [32], who also combine multi-level features to perform vehicle detection.”
From MME-YOLO: Multi-Sensor Multi-Level Enhanced YOLO for Robust Vehicle Detection in Traffic Surveillance (2020) · cites this paper as [32]
Cites
Deep Residual Learning for Image Recognition
He, K., Zhang, X., Ren, S., & Sun, J. (2016) · 2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR) · doi:10.1109/cvpr.2016.90 · Cited by 235,474 · cited 5× · see citing works
“Particularly, with the advent of deep networks on image classification task such as [11]–[14],”
BackgroundInfluentialImageNet classification with deep convolutional neural networks
Krizhevsky, A., Sutskever, I., & Hinton, G. E. (2017) · Communications of the ACM · doi:10.1145/3065386 · Cited by 112,155 · cited 3× · see citing works
“MIO-TCD is used to evaluate classification algorithms (i.e., AlexNet [11], InceptionV3 [34], ResNet50 [14], VGG-19 [12], Xception [35], and DenseNet [36]) and detection algorithms (i.e., Faster R-CNN [17], SSD-300, SSD-512 [22], YOLO [21] and YOLOv2 [23]), and achieves superior performance.”
BackgroundVery Deep Convolutional Networks for Large-Scale Image Recognition
Simonyan, K., & Zisserman, A. (2014) · arXiv · doi:10.48550/arxiv.1409.1556 · Cited by 74,248 · cited 2× · see citing works
“MIO-TCD is used to evaluate classification algorithms (i.e., AlexNet [11], InceptionV3 [34], ResNet50 [14], VGG-19 [12], Xception [35], and DenseNet [36]) and detection algorithms (i.e., Faster R-CNN [17], SSD-300, SSD-512 [22], YOLO [21] and YOLOv2 [23]), and achieves superior performance.”
Faster R-CNN: Towards Real-Time Object Detection with Region Proposal Networks
Ren, S., He, K., Girshick, R., & Sun, J. (2017) · IEEE Transactions on Pattern Analysis and Machine Intelligence · doi:10.1109/tpami.2016.2577031 · Cited by 56,035 · see citing works
You Only Look Once: Unified, Real-Time Object Detection
Redmon, J., Divvala, S., Girshick, R., & Farhadi, A. (2016) · 2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR) · doi:10.1109/cvpr.2016.91 · Cited by 53,803 · cited 2× · see citing works
“, Faster R-CNN [17], SSD-300, SSD-512 [22], YOLO [21] and YOLOv2 [23]), and achieves superior performance.”
BackgroundDensely Connected Convolutional Networks
Huang, G., Liu, Z., Van Der Maaten, L., & Weinberger, K. Q. (2017) · 2017 IEEE Conference on Computer Vision and Pattern Recognition (CVPR) · doi:10.1109/cvpr.2017.243 · Cited by 48,566 · cited 2× · see citing works
“MIO-TCD is used to evaluate classification algorithms (i.e., AlexNet [11], InceptionV3 [34], ResNet50 [14], VGG-19 [12], Xception [35], and DenseNet [36]) and detection algorithms (i.e., Faster R-CNN [17], SSD-300, SSD-512 [22], YOLO [21] and YOLOv2 [23]), and achieves superior performance.”
Going deeper with convolutions
Szegedy, C., Liu, W., Jia, Y., Sermanet, P., Reed, S., Anguelov, D., Erhan, D., Vanhoucke, V., & Rabinovich, A. (2015) · 2015 IEEE Conference on Computer Vision and Pattern Recognition (CVPR) · doi:10.1109/cvpr.2015.7298594 · Cited by 48,562 · see citing works
Rich Feature Hierarchies for Accurate Object Detection and Semantic Segmentation
Girshick, R., Donahue, J., Darrell, T., & Malik, J. (2014) · 2014 IEEE Conference on Computer Vision and Pattern Recognition · doi:10.1109/cvpr.2014.81 · Cited by 32,684 · cited 12× · see citing works
“[60] P. Li, X. Chen, and S. Shen, ‘‘Stereo R-CNN based 3D object detection for autonomous driving,’’ in Proc.”
MethodInfluentialRethinking the Inception Architecture for Computer Vision
Szegedy, C., Vanhoucke, V., Ioffe, S., Shlens, J., & Wojna, Z. (2016) · 2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR) · doi:10.1109/cvpr.2016.308 · Cited by 32,138 · cited 2× · see citing works
“MIO-TCD is used to evaluate classification algorithms (i.e., AlexNet [11], InceptionV3 [34], ResNet50 [14], VGG-19 [12], Xception [35], and DenseNet [36]) and detection algorithms (i.e., Faster R-CNN [17], SSD-300, SSD-512 [22], YOLO [21] and YOLOv2 [23]), and achieves superior performance.”
Feature Pyramid Networks for Object Detection
Lin, T.-Y., Dollar, P., Girshick, R., He, K., Hariharan, B., & Belongie, S. (2017) · 2017 IEEE Conference on Computer Vision and Pattern Recognition (CVPR) · doi:10.1109/cvpr.2017.106 · Cited by 30,939 · cited 1× · see citing works
“Inspired by [25], a concept of feature pyramid is introduced for the feature maps extracted from the multiple scales, which is similar to [22], [24].”
MethodFast R-CNNPreprint
Girshick, R. B. (2015) · arXiv (Cornell University) · doi:10.48550/arxiv.1504.08083 · Cited by 29,416 · cited 12× · see citing works
“[60] P. Li, X. Chen, and S. Shen, ‘‘Stereo R-CNN based 3D object detection for autonomous driving,’’ in Proc.”
MethodInfluentialBatch Normalization: Accelerating Deep Network Training by Reducing Internal Covariate Shift
Ioffe, S., & Szegedy, C. (2015) · arXiv · doi:10.48550/arxiv.1502.03167 · Cited by 27,818 · cited 1× · see citing works
“A batch normalization layer [50] is added on all of the convolutional layers to help regularize the network.”
BackgroundSSD: Single Shot MultiBox Detector
Liu, W., Anguelov, D., Erhan, D., Szegedy, C., Reed, S., Fu, C.-Y., & Berg, A. C. (2016) · Lecture Notes in Computer Science · doi:10.1007/978-3-319-46448-0_2 · Cited by 22,128 · cited 4× · see citing works
“Inspired by [25], a concept of feature pyramid is introduced for the feature maps extracted from the multiple scales, which is similar to [22], [24].”
BackgroundMethodInfluentialXception: Deep Learning with Depthwise Separable Convolutions
Chollet, F. (2017) · 2017 IEEE Conference on Computer Vision and Pattern Recognition (CVPR) · doi:10.1109/cvpr.2017.195 · Cited by 20,380 · cited 3× · see citing works
“MIO-TCD is used to evaluate classification algorithms (i.e., AlexNet [11], InceptionV3 [34], ResNet50 [14], VGG-19 [12], Xception [35], and DenseNet [36]) and detection algorithms (i.e., Faster R-CNN [17], SSD-300, SSD-512 [22], YOLO [21] and YOLOv2 [23]), and achieves superior performance.”
InfluentialYOLO9000: Better, Faster, Stronger
Redmon, J., & Farhadi, A. (2017) · 2017 IEEE Conference on Computer Vision and Pattern Recognition (CVPR) · doi:10.1109/cvpr.2017.690 · Cited by 19,892 · cited 8× · see citing works
“Motivated by [23], to make better predictions, we use k-means clustering on the ground truth boxes of our training dataset to find the most suitable bounding box priors.”
BackgroundMethodInfluentialYOLOv3: An Incremental Improvement
Redmon, J., & Farhadi, A. (2018) · arXiv · doi:10.48550/arxiv.1804.02767 · Cited by 17,141 · cited 7× · see citing works
“object locations and classify their corresponding categories [21]–[24].”
BackgroundMethodInfluentialAre we ready for autonomous driving? The KITTI vision benchmark suite
Geiger, A., Lenz, P., & Urtasun, R. (2012) · 2012 IEEE Conference on Computer Vision and Pattern Recognition · doi:10.1109/cvpr.2012.6248074 · Cited by 14,897 · cited 4× · see citing works
“Over the past decade, some challenging vehicle benchmarks [8]–[10] have been put forward for evaluation of various detection algorithms.”
BackgroundMethodInfluentialAggregated Residual Transformations for Deep Neural Networks
Xie, S., Girshick, R., Dollar, P., Tu, Z., & He, K. (2017) · 2017 IEEE Conference on Computer Vision and Pattern Recognition (CVPR) · doi:10.1109/cvpr.2017.634 · Cited by 12,465 · cited 1× · see citing works
“A multi-branch residual network, named ResNeXt, is designed by repeating a certain building block which aggregates a set of transformations with the same topology [44].”
BackgroundBackpropagation Applied to Handwritten Zip Code Recognition
LeCun, Y., Boser, B., Denker, J. S., Henderson, D., Howard, R. E., Hubbard, W., & Jackel, L. D. (1989) · Neural Computation · doi:10.1162/neco.1989.1.4.541 · Cited by 11,891 · cited 2× · see citing works
“C. DEEP NETS TRAINING Our fast vehicle detection network CMNet can be trained end-to-end by stochastic gradient descent (SGD) [51].”
MethodIdentity mappings in deep residual networks
He, K., Zhang, X., Ren, S., & Sun, J. (2016) · Lecture Notes in Computer Science · doi:10.1007/978-3-319-46493-0_38 · Cited by 10,741 · cited 1× · see citing works
“Since residual networks [14], [38] have achieved great success, recently, various deep fusion residual network architectures have been designed to enhance the representation power”
BackgroundZagoruyko, S., & Komodakis, N. (2016) · Procedings of the British Machine Vision Conference 2016 · doi:10.5244/c.30.87 · Cited by 7,244 · cited 1× · see citing works
“A novel architecture based on residual networks, which is called wide residual networks (WRNs), is present in [39] to address the problem of slow training.”
BackgroundInception-v4, Inception-ResNet and the Impact of Residual Connections on Learning
Szegedy, C., Ioffe, S., Vanhoucke, V., & Alemi, A. (2017) · Proceedings of the AAAI Conference on Artificial Intelligence · doi:10.1609/aaai.v31i1.11231 · Cited by 4,636 · cited 1× · see citing works
“There is empirical evidence that combining the Inception architecture with residual connections can outperform similarly expensive Inception networks without residual connections [43].”
BackgroundMulti-view 3D Object Detection Network for Autonomous Driving
Chen, X., Ma, H., Wan, J., Li, B., & Xia, T. (2017) · 2017 IEEE Conference on Computer Vision and Pattern Recognition (CVPR) · doi:10.1109/cvpr.2017.691 · Cited by 3,450 · see citing works
PointRCNN: 3D Object Proposal Generation and Detection From Point Cloud
Shi, S., Wang, X., & Li, H. (2019) · 2019 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) · doi:10.1109/cvpr.2019.00086 · Cited by 3,093 · see citing works
Spatial Pyramid Pooling in Deep Convolutional Networks for Visual Recognition
He, K., Zhang, X., Ren, S., & Sun, J. (2014) · Lecture Notes in Computer Science · doi:10.1007/978-3-319-10578-9_23 · Cited by 3,034 · cited 2× · see citing works
“However, the challenge is that two-stage based algorithms, such as R-CNN [18], SPP-Net [19], Fast R-CNN [20], and Faster R-CNN [17], are always too”
MethodEdge Boxes: Locating Object Proposals from Edges
Zitnick, C. L., & Dollár, P. (2014) · Lecture Notes in Computer Science · doi:10.1007/978-3-319-10602-1_26 · Cited by 2,656 · cited 1× · see citing works
“For two-stage based approaches, Selective Search [15], EdgeBoxes [16], Region Proposal Network (RPN) [17] are often utilized to generate region proposals for detection.”
MethodJoint 3D Proposal Generation and Object Detection from View Aggregation
Ku, J., Mozifian, M., Lee, J., Harakeh, A., & Waslander, S. L. (2018) · 2018 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS) · doi:10.1109/iros.2018.8594049 · Cited by 1,697 · see citing works
Single-Shot Refinement Neural Network for Object Detection
Zhang, S., Wen, L., Bian, X., Lei, Z., & Li, S. Z. (2018) · 2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition · doi:10.1109/cvpr.2018.00442 · Cited by 1,678 · see citing works
DSSD : Deconvolutional Single Shot Detector
Fu, C.-Y., Liu, W., Ranga, A., Tyagi, A., & Berg, A. C. (2017) · arXiv · doi:10.48550/arxiv.1701.06659 · Cited by 1,624 · see citing works
Scalable Object Detection Using Deep Neural Networks
Erhan, D., Szegedy, C., Toshev, A., & Anguelov, D. (2014) · 2014 IEEE Conference on Computer Vision and Pattern Recognition · doi:10.1109/cvpr.2014.276 · Cited by 1,161 · cited 2× · see citing works
“Note that for each ground truth object we only assign one bounding box prior, which is used by MultiBox [52], to reduce the repeat boxes for one vehicle object.”
MethodMonocular 3D Object Detection for Autonomous Driving
Chen, X., Kundu, K., Zhang, Z., Ma, H., Fidler, S., & Urtasun, R. (2016) · 2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR) · doi:10.1109/cvpr.2016.236 · Cited by 1,019 · see citing works
A large-scale car dataset for fine-grained categorization and verification
Yang, L., Luo, P., Loy, C. C., & Tang, X. (2015) · IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings - IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings · doi:10.1109/cvpr.2015.7299023 · Cited by 867 · cited 1× · see citing works
“DAVE, which is trained on CompCars dataset [27], can achieve significant improvements compared with other existing algorithms on self-collected urban traffic surveillance dataset, PASCAL VOC 2007 car dataset [28] and LISA 2010 dataset [5].”
MethodUA-DETRAC: A New Benchmark and Protocol for Multi-Object Detection and Tracking
Wen, L., Du, D., Cai, Z., Lei, Z., Chang, M.-C., Qi, H., Lim, J., Yang, M.-H., & Lyu, S. (2020) · Computer Vision and Image Understanding · doi:10.1016/j.cviu.2020.102907 · Cited by 804 · cited 1× · see citing works
“work on a quarter high definition (960 × 540 pixels) UA-DETRAC dataset [9], [10], which contains 82082 realworld vehicle images over four categories of car, bus, van, and others.”
BackgroundResnet in Resnet: Generalizing Residual ArchitecturesPreprint
Targ, S., Almeida, D., & Lyman, K. (2016) · arXiv (Cornell University) · doi:10.48550/arxiv.1603.08029 · Cited by 701 · cited 2× · see citing works
“Available: https://arxiv.org/abs/1605.07146 [40] S. Targ, D. Almeida, and K. Lyman, ‘‘Resnet in resnet: Generalizing residual architectures,’’ 2016, arXiv:1603.08029.”
BackgroundInfluentialYang, F., Choi, W., & Lin, Y. (2016) · 2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR) · doi:10.1109/cvpr.2016.234 · Cited by 608 · see citing works
Self-citationStereo R-CNN Based 3D Object Detection for Autonomous Driving
Li, P., Chen, X., & Shen, S. (2019) · IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings - IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings · doi:10.1109/cvpr.2019.00783 · Cited by 597 · see citing works
End-to-End People Detection in Crowded Scenes
Stewart, R., Andriluka, M., & Ng, A. Y. (2016) · 2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR) · doi:10.1109/cvpr.2016.255 · Cited by 539 · see citing works
Residual Networks of Residual Networks: Multilevel Residual Networks
Zhang, K., Sun, M., Han, T. X., Yuan, X., Guo, L., & Liu, T. (2017) · IEEE Transactions on Circuits and Systems for Video Technology · doi:10.1109/tcsvt.2017.2654543 · Cited by 398 · cited 1× · see citing works
“In [41], an enhanced version of residual networks, Residual networks of Residual networks (RoR), adds level-wise shortcut connections upon conventional residual networks to dig the optimization ability of residual networks.”
BackgroundA General Active-Learning Framework for On-Road Vehicle Recognition and Tracking
Sivaraman, S., & Trivedi, M. (2010) · IEEE transactions on intelligent transportation systems (Print) · doi:10.1109/tits.2010.2040177 · Cited by 389 · cited 1× · see citing works
“DAVE, which is trained on CompCars dataset [27], can achieve significant improvements compared with other existing algorithms on self-collected urban traffic surveillance dataset, PASCAL VOC 2007 car dataset [28] and LISA 2010 dataset [5].”
MethodSINet: A Scale-Insensitive Convolutional Neural Network for Fast Vehicle Detection
Hu, X., Xu, X., Xiao, Y., Chen, H., He, S., Qin, J., & Heng, P.-A. (2019) · IEEE Transactions on Intelligent Transportation Systems · doi:10.1109/tits.2018.2838132 · Cited by 341 · cited 1× · see citing works
“A scale-insensitive convolutional neural network (SINet) is presented in [3], which trains on KITTI benchmark and a new highway dataset, for detecting vehicles with a large variance of scales.”
BackgroundNeural Networks for Pattern Recognition
Kothari, S. C., & Oh, H. (1993) · Advances in computers · doi:10.1016/s0065-2458(08)60404-0 · Cited by 316 · cited 1× · see citing works
“The operation Ft (xt ) + xt represents a shortcut connection [14], [48], [49].”
BackgroundSubcategory-Aware Convolutional Neural Networks for Object Proposals and Detection
Yu, X., Choi, W., Lin, Y., & Savarese, S. (2017) · IEEE Winter Conference on Applications of Computer Vision · doi:10.1109/wacv.2017.108 · Cited by 303 · see citing works
Cloud-Based Cyber-Physical Intrusion Detection for Vehicles Using Deep Learning
Loukas, G., Vuong, T. P., Heartfield, R., Sakellari, G., Yoon, Y., & Gan, D. (2017) · IEEE Access · doi:10.1109/access.2017.2782159 · Cited by 278 · see citing works
Structure Inference Net: Object Detection Using Scene-Level Context and Instance-Level Relationships
Liu, Y., Wang, R., Shan, S., & Chen, X. (2018) · IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings - IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings · doi:10.1109/cvpr.2018.00730 · Cited by 263 · see citing works
Symmetrical SURF and Its Applications to Vehicle Detection and Vehicle Make and Model Recognition
Hsieh, J.-W., Chen, L.-C., & Chen, D.-Y. (2014) · IEEE Transactions on Intelligent Transportation Systems · doi:10.1109/tits.2013.2294646 · Cited by 200 · see citing works
Nejad, H. T. N., Takeuchi, A., Mita, S., & McAllester, D. (2012) · IEEE Transactions on Intelligent Transportation Systems · doi:10.1109/tits.2012.2187894 · Cited by 160 · cited 1× · see citing works
“Among many vehicle detection algorithms, deep learning-based methods have attracted great attention and been widely used [1]–[7].”
BackgroundMIO-TCD: A New Benchmark Dataset for Vehicle Classification and Localization
Luo, Z., Branchaud-Charron, F., Lemaire, C., Konrad, J., Li, S., Mishra, A. K., Achkar, A., Eichel, J. A., & Jodoin, P. (2018) · IEEE Transactions on Image Processing · doi:10.1109/tip.2018.2848705 · Cited by 152 · cited 1× · see citing works
“It is noteworthy that a new benchmark dataset contains 786702 labeled images is introduced in [33], which is denoted as MIO-TCD, for vehicle classification and localization.”
BackgroundArbitrary-Oriented Vehicle Detection in Aerial Imagery with Single Convolutional Neural Networks
Tang, T.-Y., Zhou, S.-L., Deng, Z., Lei, L., & Zou, H. (2017) · Remote Sensing · doi:10.3390/rs9111170 · Cited by 123 · cited 1× · see citing works
“To improve the vehicle detection performance in aerial images, especially for vehicles in multiple directions, an enhanced version of SSD, which is called as Oriented_SSD, is proposed to directly produce arbitrarilyoriented detection results [37].”
MethodFast Detection of Multiple Objects in Traffic Scenes With a Common Detection Framework
Hu, Q., Paisitkriangkrai, S., Shen, C., van den Hengel, A., & Porikli, F. (2015) · IEEE Transactions on Intelligent Transportation Systems · doi:10.1109/tits.2015.2496795 · Cited by 123 · see citing works
Evolving boxes for fast vehicle detection
Wang, L., Lu, Y., Wang, H., Zheng, Y., Ye, H., & Xue, X. (2017) · IEEE International Conference on Multimedia and Expo (ICME) · doi:10.1109/icme.2017.8019461 · Cited by 118 · cited 1× · see citing works
“Recently, to refine the bounding boxes of vehicles under different featuremaps, a framework named Evolving Boxes is presented [30].”
BackgroundUA-DETRAC 2017: Report of AVSS2017 & IWT4S Challenge on Advanced Traffic Monitoring
Lyu, S., Chang, M., Du, D., Wen, L., Qi, H., Li, Y., Wei, Y., Ke, L., Hu, T., Coco, M. D., et al. (2017) · IEEE International Conference on Advanced Video and Signal Based Surveillance (AVSS) · doi:10.1109/avss.2017.8078560 · Cited by 92 · cited 2× · see citing works
“Over the past decade, some challenging vehicle benchmarks [8]–[10] have been put forward for evaluation of various detection algorithms.”
BackgroundMulti-Task Vehicle Detection With Region-of-Interest Voting
Chu, W., Liu, Y., Shen, C., Cai, D., & Hua, X. (2017) · IEEE Transactions on Image Processing · doi:10.1109/tip.2017.2762591 · Cited by 89 · cited 2× · see citing works
“Among many vehicle detection algorithms, deep learning-based methods have attracted great attention and been widely used [1]–[7].”
BackgroundMethodDeeply-Fused NetsPreprint
Wang, J., Wei, Z., Zhang, T., & Zeng, W. (2016) · arXiv (Cornell University) · doi:10.48550/arxiv.1605.07716 · Cited by 71 · cited 1× · see citing works
“Inspired by the idea of deep fusion [45], as well as fusion methods between residual blocks [46], [47], this paper propose a novel connect-and-merge convolutional neural network to fast and accurate detect vehicles with various scales and types in urban traffic surveillance.”
MethodMin, W., Fan, M., Guo, X., & Han, Q. (2017) · IEEE Transactions on Intelligent Transportation Systems · doi:10.1109/tits.2017.2756989 · Cited by 71 · cited 1× · see citing works
“In [29], an improved ViBe algorithm is proposed to track multiple vehicles in various traffic scenes,”
MethodDeep Convolutional Neural Networks with Merge-and-Run Mappings
Zhao, L., Li, M., Meng, D.-P., Li, X., Zhang, Z., Zhuang, Y., Tu, Z.-W., & Wang, J. (2016) · International Joint Conference on Artificial Intelligence · doi:10.24963/ijcai.2018/440 · Cited by 70 · cited 2× · see citing works
“Inspired by the idea of deep fusion [45], as well as fusion methods between residual blocks [46], [47], this paper propose a novel connect-and-merge convolutional neural network to fast and accurate detect vehicles with various scales and types in urban traffic surveillance.”
MethodCao, W., Yuan, J.-H., He, Z., Zhang, Z., & He, Z.-Q. (2018) · IEEE Access · doi:10.1109/access.2018.2795798 · Cited by 57 · cited 1× · see citing works
“In [32], a deep neural network is proposed by utilizing ideas of knowledge guided training and predicted regions of interest, to significantly improve vehicle detection speed while ensuring accuracy.”
BackgroundMulti-Residual Networks: Improving the Speed and Accuracy of Residual NetworksPreprint
Abdi, M., & Nahavandi, S. (2016) · arXiv (Cornell University) · doi:10.48550/arxiv.1609.05672 · Cited by 54 · cited 1× · see citing works
“blocks, a multi-residual network is proposed [42], which significantly improves the speed and accuracy compared with very deep residual networks.”
BackgroundRefineNet: Refining Object Detectors for Autonomous Driving
Rajaram, R. N., Ohn-Bar, E., & Trivedi, M. (2016) · IEEE Transactions on Intelligent Vehicles · doi:10.1109/tiv.2017.2695896 · Cited by 51 · see citing works
Fast Automatic Vehicle Annotation for Urban Traffic Surveillance
Zhou, Y., Liu, L., Shao, L., & Mellor, M. (2018) · IEEE transactions on intelligent transportation systems (Print) · doi:10.1109/tits.2017.2740303 · Cited by 46 · cited 3× · see citing works
“Inspired by [8], [26], we employ the establishedmean average precision (mAP) as accuracy and frames per second (FPS) FIGURE 6.”
MethodScene-Adaptive Vehicle Detection Algorithm Based on a Composite Deep Structure
Cai, Y.-F., Wang, H., Zheng, Z., & Sun, X.-Q. (2017) · IEEE Access · doi:10.1109/access.2017.2756081 · Cited by 24 · cited 1× · see citing works
“vehicle detection results in different target scenes, a sceneadaptive algorithm based on a composite deep structure is proposed in [2], which performs better than both existing scene-adaptive based and machine learning based methods.”
MethodA Graph-Based Vehicle Proposal Location and Detection Algorithm
Yuan, X., Su, S., & Chen, H.-J. (2017) · IEEE transactions on intelligent transportation systems (Print) · doi:10.1109/tits.2017.2676182 · Cited by 20 · cited 1× · see citing works
“A vehicle proposal location method is employed in [31], which involves a flexible bounding-box”
Method
Flagged first, then by citations
5 of the 66 references couldn’t be matched to a known paper.
Cited by
Sorted by citations
Qin, B., & Li, D. (2020) · Sensors · doi:10.3390/s20185236 · Cited by 300 · see citing works
“Upscaling low-quality (low-resolution, blurred) input images to produce high-quality feature maps to improve classification performance is one of the most popular ways for low-quality image classification or object detection [53–57].”
BackgroundGe, L.-F., Dan, D.-H., & Li, H. (2020) · Structural Control & Health Monitoring · doi:10.1002/stc.2636 · Cited by 115 · see citing works
“Zhang et al.32 classified and located vehicles on a bridge based on Faster-RCNN.”
MethodZhu, J., Li, X., Jin, P., Xu, Q., Sun, Z., & Song, X. (2020) · Sensors · doi:10.3390/s21010027 · Cited by 51 · see citing works
“Our works achieve better performance compared with recent robust detectors, such as [31] and [32], who also combine multilevel features to perform vehicle detection.”
BackgroundZarei, N., Moallem, P., & Shams, M. (2022) · IEEE Access · doi:10.1109/access.2022.3221942 · Cited by 46 · see citing works
“A convolutional neural network (CMNet) was proposed for fast vehicle detection in complex scenes [40].”
MethodA Robust Vehicle Detection Scheme for Intelligent Traffic Surveillance Systems in Smart Cities
Wang, Z., Huang, J.-F., Xiong, N., Zhou, X.-P., Lin, X., & Ward, T. L. (2020) · IEEE Access · doi:10.1109/access.2020.3012995 · Cited by 39 · see citing works
“[16] proposed a connect-and-merge convolutional neural network (CMNet) for fast vehicle detection in urban traffic surveillance.”
BackgroundAccurate and efficient vehicle detection framework based on SSD algorithm
Zhao, M., Zhong, Y., Sun, D.-H., & Chen, Y. (2021) · IET Image Processing · doi:10.1049/ipr2.12297 · Cited by 35 · see citing works
“CMNet [22] uses YOLO V3 [7] as the baseline algorithm and employs a carefully designed Connetct-and-Merge Block, which can effectively improve the flow of information in CNN.”
MethodCNN Convolutional layer optimisation based on quantum evolutionary algorithm
Lu, T.-C. (2020) · Connection Science · doi:10.1080/09540091.2020.1841111 · Cited by 33 · see citing works
“…have made breakthrough developments in terms of image recognition (Bae et al., 2020; Das et al., 2019; Li et al., 2020; Liang et al., 2018; Liu et al., 2020; Nakazawa & Kulkarni, 2018; Rezaee et al., 2018; Srivastava & Biswas, 2020; Yamashita et al., 2018; Zhang et al., 2019; Zhu et al., 2020).”
BackgroundVehicle tracking with Kalman filter using online situation assessment
Baradaran-Khalkhali, M., Vahedian, A., & Yazdi, H. (2020) · Robotics Auton. Syst. · doi:10.1016/j.robot.2020.103596 · Cited by 27
YOLO-Anti: YOLO-based counterattack model for unseen congested object detection
Wang, K., & Liu, M. (2022) · Pattern Recognition · doi:10.1016/j.patcog.2022.108814 · Cited by 20 · see citing works
Guo, F., Wang, Y., & Qian, Y. (2022) · Journal of Industrial Information Integration · doi:10.1016/j.jii.2022.100427 · Cited by 19 · see citing works
Multiview image generation for vehicle reidentification
Zhang, F., Ma, Y., Yuan, G., Zhang, H., & Ren, J. (2021) · Applied intelligence (Boston) · doi:10.1007/s10489-020-02171-8 · Cited by 18 · see citing works
Self-citationVu, V. Q., Tran, M.-Q., Amer, M., Khatiwada, M., Ghoneim, S., & Elsisi, M. (2023) · Inf. · doi:10.3390/info14070379 · Cited by 14 · see citing works
Sabir, M., Mehmood, I., Alsaggaf, W., Khairullah, E., Alhuraiji, S., Alghamdi, A., & El-Latif, A. A. A. (2022) · Computers Materials & Continua · doi:10.32604/cmc.2022.017865 · Cited by 14 · see citing works
Image-to-image domain adaptation for vehicle re-identification
Zhang, F., Zhang, L., Zhang, H., & Ma, Y. (2023) · Multimedia tools and applications · doi:10.1007/s11042-023-14839-7 · Cited by 9 · see citing works
“…image-to-image translation using a vehicle transfer generative adversarial network based on the reconstructed HR images, after that, we use CMNet-50 [48] with local max pooling as a feature extractor to learn an enhanced vehicle representations free from style variations to facilitate vehicle ReID.”
MethodSelf-citationReal-time vehicle detection using segmentation-based detection network and trajectory prediction
Zarei, N., Moallem, P., & Shams, M. (2023) · IET Computer Vision · doi:10.1049/cvi2.12236 · Cited by 7 · see citing works
Connie, T., Wee, R., & Goh, M. K. O. (2024) · International Journal on Advanced Science, Engineering and Information Technology · doi:10.18517/ijaseit.14.6.12376 · Cited by 5
“Zhang et al. [27] proposed a method for detecting fast vehicles by using a method called connect-and-merge convolutional neural network CMNet.”
Multi-Class Vehicle Detection Using VDnet in Heterogeneous Traffic
Deshmukh, P., Rayasam, K. C., Sahoo, U. K., Das, S. K., & Majhi, S. (2024) · IEEE transactions on intelligent transportation systems (Print) · doi:10.1109/tits.2024.3476122 · Cited by 5 · see citing works
Vehicle type classification using graph ant colony optimizer based stack autoencoder model
Rani, B., Rao, M. V., Patra, R. K., Srinivas, K., & Madhukar, G. (2022) · Multimedia tools and applications · doi:10.1007/s11042-021-11508-5 · Cited by 4
“The image based techniques gained more attention among the researchers in vehicle type classification, due to the use of the traffic surveillance cameras [7, 31].”
MethodSudhaS, K., & Aji, S. (2025) · Knowledge-Based Systems · doi:10.1016/j.knosys.2025.114067 · Cited by 3 · see citing works
Garbage Classification using Faster R-CNN
Ma'rifah, P. N., Sarosa, M., & Rohadi, E. (2023) · 2023 International Conference on Electrical and Information Technology (IEIT) · doi:10.1109/ieit59852.2023.10335519 · Cited by 3
“…based on deep learning have good performance, convolutional neural networks (CNNs) are one of the most common and widely used deep learning methods [26][27][28] because most of the deep convolutional neural network architectures show superior capabilities in studying hierarchical and…”
MethodOptimizing Vehicle Classification using YOLOv8: Strategies to Counter Overfitting
Jha, A., Saxena, I., Kadian, K., Indu, Sharma, V., & Singh, S. (2023) · 2023 3rd International Conference on Advancement in Electronics & Communication Engineering (AECE) · doi:10.1109/aece59614.2023.10428452 · Cited by 2 · see citing works
“They also employed a fresh regularization method created especially for automobile object identification and classification tasks [13].”
MethodTraffic Risk Modelling and Analysis under Airport‐Like Simulation Environment
Sheng, Y., Feng, R., & Biancardo, S. A. (2022) · Journal of Advanced Transportation · doi:10.1155/2022/7907435 · Cited by 2
“Zhang et al. proposed a connect-and-merge residual network to extract simplified but robust vehicle features under complex roadway scene interference [9].”
Lv, Y., Ning, M., Zhou, F., Lv, P., Zhang, P.-Y., & Wang, J. (2024) · Journal of Supercomputing · doi:10.1007/s11227-024-06455-5 · Cited by 1
A Review on Object Detection in Traffic Scenarios Using Deep Learning Models
R, K., R, D., & Purnachand, K. (2025) · 2025 International Conference on Visual Analytics and Data Visualization (ICVADV) · doi:10.1109/icvadv63329.2025.10961569
“UA-DETRAC dataset was used in [24] and [27].”
InfluentialERODE: Error Resilient Object DetEction by Recovering Bounding Box and Class Information
Valpreda, E., Palumbo, G., Caon, M., Masera, G., & Martina, M. (2023) · Conference on Ph.D. Research in Microelectronics and Electronics · doi:10.1109/prime58259.2023.10161894
“They are increasingly used in safety-critical applications such as autonomous driving, or traffic surveillance [1], [2].”
BackgroundPedestrian-Vehicle Detection Model Based on Optimized YOLOv3
Tang, K., Yan, F., & Liu, W.-B. (2020) · ACM Cloud and Autonomic Computing Conference · doi:10.1109/cac51589.2020.9327192
“Regarding the problem of obtaining more semantic information, reference [28] proposed a CMNet that connects and merges residual networks (CMRN) for feature extraction.”
Background
- Authors
- Zhang, F., Yang, F., Li, C., & Yuan, G.
- Affiliations
- China University of Mining and Technology - Beijing (CN)
- Guilin University of Electronic Technology (CN)
- Venue
- IEEE Access
- Published
- 2019
- Volume
- 7
- Pages
- 72660-72671
- ISSN
- 2169-3536 (ISSN-L)
- OpenAlex
- W2947050565
- Cited by
- 29
- References
- 66
- Citations by year
- 2025: 2 · 2024: 3 · 2023: 7 · 2022: 6 · 2021: 4 · 2020: 7
- FWCI
- 1.90
- Type
- journal article
- Language
- en
- Open access
- gold
- doi.org
- Locations
- IEEE Access (publishedVersion)
- DOAJ (DOAJ: Directory of Open Access Journals) (submittedVersion, cc-by-sa)
- Funders
- Fundamental Research Funds for the Central Universities (2016QJ04)
- Yue Qi Young Scholar Project of CUMTB
- National Natural Science Foundation of China (61601466)
- Guangxi Key Laboratory of Trusted Software (KX201613)
- Topics
- Physical Sciences › Computer Science › Computer Vision and Pattern Recognition › Advanced Neural Network Applications
- Physical Sciences › Computer Science › Computer Vision and Pattern Recognition › Video Surveillance and Tracking Methods
- Physical Sciences › Engineering › Automotive Engineering › Autonomous Vehicle Technology and Safety
- Keywords
- vehicle detection, urban traffic surveillance, ResNet, feature extraction, semantic information, real-time detection, KITTI dataset
About these checks
Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.
Statistics and arithmetic are recomputed, not judged.
Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.
What counts as a claim ›How we measure confidence ›How we check this ›
Are you an author, or do you see a problem with this page? Request a review
