Smart City Gnosys

Smart city article details

Title	Compress Yolov5 Via Convolution Filters Reconstruction For The Application In Smart City Governance
ID_Doc	15344
Authors	Wang S.; Wang Y.; Lan W.; Yang Y.; Xin C.; Yu Y.; Ma C.
Year	2025
Published	2025 IEEE 17th International Conference on Computer Research and Development, ICCRD 2025
DOI	http://dx.doi.org/10.1109/ICCRD64588.2025.10962968
Abstract	In the realm of smart city governance, object detection techniques have become increasingly integral. With the YOLO architecture emerges, it has become a prominent choice for detection due to its efficiency and accuracy. Despite such advantages, the computational demands of YOLO pose challenges for deployment on platforms with limited resources in many real applications. To address this, we introduce an compression methodology tailored for the YOLO model, aiming at enhancing the operational efficiency and practical applicability within the context of urban management. Specifically, for a convolutional layer, we reconstruct the convolution filters based on a set of filter bases, whose amount is much lower than original number of convolution filters. Using YOLOv5 as a case study, we implement the method by replacing the convolutional layer in C3 modules with our proposed new filters. The empirical evaluation of our method, conducted on three custom datasets, demonstrates that the compressed YOLOv5 model attains a substantial reduction in both parameter and FLOPs count, while maintaining the detection accuracy. This verifies the effectiveness of proposed method in practical applications. In addition, we also explore the specific impact of different compression strategies and parameter setting on model performance. The findings offer valuable insights into optimizing the balance between model scale and detection accuracy, contributing to the development of efficient object detection solution for resource constrained devices in smart city governance. © 2025 IEEE.
Author Keywords	model compression; object detection; smart city

Similar Articles

Id	Similarity	Authors	Title	Published
40259	0.876	Yang Q.; Song L.; Zhou H.	Opportunities And Challenges Of Yolo -World In Smart City Surveillance	Proceedings - 2024 2nd International Conference on Mechatronics, IoT and Industrial Informatics, ICMIII 2024 (2024)
54356	0.874	Padilla Carrasco D.; Rashwan H.A.; Garcia M.A.; Puig D.	T-Yolo: Tiny Vehicle Detection Based On Yolo And Multi-Scale Convolutional Neural Networks	IEEE Access, 11 (2023)
62125	0.871	Song F.; Li P.	Yolov5-Ms: Real-Time Multi-Surveillance Pedestrian Target Detection Model For Smart Cities	Biomimetics, 8, 6 (2023)
38160	0.865	Wang Y.; Liu X.; Wang R.	Multi-Class Object Detection In Urban Scenes Based On Deep Learning	ISPRS Annals of the Photogrammetry, Remote Sensing and Spatial Information Sciences, 10, 3 (2024)
44417	0.864	Kurinaah E.; Hall M.; Viera N.; Smith K.; Yu Y.; Shen X.	Real-Time Parking Lot Monitoring For Smart Cities: A Cnn-Based Approach Using Yolo And Rtsp-Compatible Cameras	Proceedings of SPIE - The International Society for Optical Engineering, 13465 (2025)
15079	0.864	Khalfaoui A.; Badri A.; Mourabit I.E.L.	Comparative Study Of Yolov3 And Yolov5'S Performances For Real-Time Person Detection	2022 2nd International Conference on Innovative Research in Applied Science, Engineering and Technology, IRASET 2022 (2022)
50560	0.863	Baiat Z.E.; Baydere S.	Smart City Traffic Monitoring:Yolov7 Transfer Learning Approach For Real-Time Vehicle Detection	2023 International Conference on Smart Applications, Communications and Networking, SmartNets 2023 (2023)
765	0.861	Shokri D.; Larouche C.; Homayouni S.	A Comparative Analysis Of Multi-Label Deep Learning Classifiers For Real-Time Vehicle Detection To Support Intelligent Transportation Systems	Smart Cities, 6, 5 (2023)
30916	0.86	Dalal S.; Lilhore U.K.; Sharma N.; Arora S.; Simaiya S.; Ayadi M.; Almujally N.A.; Ksibi A.	Improving Smart Home Surveillance Through Yolo Model With Transfer Learning And Quantization For Enhanced Accuracy And Efficiency	PeerJ Computer Science, 10 (2024)
47306	0.855	Mao K.; Jin R.; Ying L.; Yao X.; Dai G.; Fang K.	Sc-Yolo: Provide Application-Level Recognition And Perception Capabilities For Smart City Industrial Cyber-Physical System	IEEE Systems Journal, 17, 4 (2023)