YOLOv5-CBAM-EfficientNet: an optimized deep learning framework for traffic sign recognition in autonomous driving
Amit Juyal1, Shuchi Bhadula2 and Sachin Sharma3
Department of Computer Science and Engineering,Graphic Era Deemed to be University, Graphic Era Hill University,Dehradun,India1
Department of Computer Science and Engineering,Graphic Era Deemed to be University,Dehradun,India2
Amity School of Engineering and Technology,Amity University, Mohali,Punjab,India3
Department of Computer Science and Engineering,Graphic Era Deemed to be University,Dehradun,India2
Amity School of Engineering and Technology,Amity University, Mohali,Punjab,India3
Corresponding Author : Amit Juyal
Recieved : 14-October-2024; Revised : 10-September-2025; Accepted : 14-September-2025
Abstract
Traffic sign detection and recognition (TSDR) is a critical feature for both self-driving cars and autonomous vehicles (AVs), as well as for modern advanced driver assistance systems (ADAS). Real-time and accurate traffic sign detection enhances safe driving capabilities of AVs and contributes to optimized traffic flow. In the context of Indian roads, challenges such as diverse sign designs, occlusion, and inconsistent lighting conditions reduce the effectiveness of standard models. Convolutional neural networks (CNNs) form the backbone of many deep learning approaches for object detection and image classification. Various CNN-based techniques have shown significant success in TSDR. However, in real-world traffic environments, deteriorate due to weather, sustain damage, or become partially obscured by rust or dust. To address these challenges, a custom dataset comprising 3,337 images of Indian traffic signs was created. Data augmentation techniques such as blurring, occlusion (cut), and grayscale conversion were employed to enhance dataset robustness. A novel hybrid deep learning framework, YOLOv5-CBAM-EfficientNet, is proposed by integrating you only look once (YOLOv5) as the backbone, EfficientNet, and the convolutional block attention module (CBAM) to improve spatial and channel-wise feature attention. The model achieved a mean average precision (mAP) of 99% while maintaining real-time performance at 45 frames per second (FPS), demonstrating its potential for use in AVs. The results indicate that the proposed model significantly enhances the perception module of AVs operating in the complex Indian traffic environment.
Keywords
Traffic sign detection, Autonomous vehicles, YOLOv5, EfficientNet, Convolutional block attention module (CBAM).
Cite this article
Juyal A, Bhadula S, Sharma S. YOLOv5-CBAM-EfficientNet: an optimized deep learning framework for traffic sign recognition in autonomous driving. International Journal of Advanced Technology and Engineering Exploration. 2025;12(130):1432-1448. DOI : 10.19101/IJATEE.2024.111101874
References
[1]
Flores-calero M, Astudillo CA, Guevara D, Maza J, Lita BS, Defaz B, et al. Traffic sign detection and recognition using YOLO object detection algorithm: a systematic review. Mathematics. 2024; 12(2):1-31.
[2]
Aldoski ZN, Koren C. Traffic sign detection and quality assessment using YOLOv8 in daytime and nighttime conditions. Sensors. 2025; 25(4):1-29.
[3]
Hosseinian SM, Mirzahossein H. Efficiency and safety of traffic networks under the effect of autonomous vehicles. Iranian Journal of Science and Technology, Transactions of Civil Engineering. 2024; 48(4):1861-85.
[4]
Prakash AJ, Sruthy S. Enhancing traffic sign recognition (TSR) by classifying deep learning models to promote road safety. Signal, Image and Video Processing. 2024; 18(5):4713-29.
[5]
Sreemathy R, Turuk M, Patil H, Shastri S, Lele N, Khurana S. Traffic sign detection and recognition using convolutional neural networks. In AI in the social and business world: a comprehensive approach 2024 (pp. 139-51). Bentham Science Publishers.
[6]
Gupta R, Kumar P, Sobti P, Kumar B, Shivam Y, Shukla P. Efficient real-time sign detection for autonomous vehical in hazy environment using deep learning models. In international conference on advanced network technologies and intelligent computing 2023 (pp. 239-57). Cham: Springer Nature Switzerland.
[7]
Le BL, Lam GH, Nguyen XV, Nguyen TM, Duong QL, Tran QD, et al. A deep learning based traffic sign detection for intelligent transportation systems. In international conference on computational data and social networks 2021 (pp. 129-37). Cham: Springer International Publishing.
[8]
Thomas S, Mathew J. Effect of color space transforms and fractional derivative on traffic sign images. In 1st international conference on trends in engineering systems and technologies (ICTEST) 2024 (pp. 1-4). IEEE.
[9]
Subhamathi AS, Sathiyapriyan C, Anand AJ, Rheem A, Arun MR. Traffic sign detection problems using convolutional neural techniques in image processing. In international conference on intelligent systems for cybersecurity (ISCS) 2024 (pp. 1-6). IEEE.
[10]
Rubio A, Demoor G, Chalmé S, Sutton-charani N, Magnier B. Sensitivity analysis of traffic sign recognition to image alteration and training data size. Information. 2024; 15(10):1-22.
[11]
CR RK, Pallavi NR. Machine learning-powered traffic sign tracking: a step-by-step framework for recognition. In second international conference on advances in information technology (ICAIT) 2024 (pp. 1-6). IEEE.
[12]
Wang Y, Wang J, Wang Q. Enhancing autonomous driving safety: a robust stacking ensemble model for traffic sign detection and recognition. Sustainability. 2024; 16(19):8597.
[13]
Mogelmose A, Trivedi MM, Moeslund TB. Vision-based traffic sign detection and analysis for intelligent driver assistance systems: perspectives and survey. IEEE Transactions on Intelligent Transportation Systems. 2012; 13(4):1484-97.
[14]
Yucong S, Shuqing G. Traffic sign recognition based on HOG feature extraction. Journal of Measurements in Engineering. 2021; 9(3):142-55.
[15]
Sugiharto A, Wirawan PW, Nugroho FA, Yudantomo SA, Pratama R, Mahe MH, et al. Comparison of SVM, random forest and KNN classification by using HOG on traffic sign detection. In 6th international conference on informatics and computational sciences (ICICoS) 2022 (pp. 60-5). IEEE.
[16]
Girshick R, Donahue J, Darrell T, Malik J. Rich feature hierarchies for accurate object detection and semantic segmentation. In proceedings of the conference on computer vision and pattern recognition 2014 (pp. 580-7). IEEE.
[17]
Redmon J, Divvala S, Girshick R, Farhadi A. You only look once: unified, real-time object detection. In proceedings of the conference on computer vision and pattern recognition 2016 (pp. 779-88). IEEE.
[18]
Wei L. SSD: single shot multibox detector. In computer vision ECCV 14th European conference proceedings part I 2016 (pp. 21-37). ACM.
[19]
Juyal A, Sharma S, Matta P. Traffic sign detection using deep learning techniques in autonomous vehicles. In international conference on innovative computing, intelligent communication and smart electrical systems (ICSES) 2021 (pp. 1-7). IEEE.
[20]
Li P, Liu C, Li T, Wang X, Zhang S, Yu D. Emdfnet: efficient multi-scale and diverse feature network for traffic sign detection. In international conference on artificial neural networks 2024 (pp. 120-36). Cham: Springer Nature Switzerland.
[21]
Zhang H, Zhang W, Wang W, Li X, Zhang A. Research on traffic sign detection algorithm based on improved SSD in complex environments. Measurement Science and Technology. 2024; 35(11):115404.
[22]
Zhu Y, Yan WQ. Traffic sign recognition based on deep learning. Multimedia Tools and Applications. 2022; 81(13):17779-91.
[23]
Kim CI, Park J, Park Y, Jung W, Lim YS. Deep learning-based real-time traffic sign recognition system for urban environments. Infrastructures. 2023; 8(2):1-19.
[24]
Alawaji K, Hedjar R, Zuair M. Traffic sign recognition using multi-task deep learning for self-driving vehicles. Sensors. 2024; 24(11):1-13.
[25]
Zhang J, Li X, Wang Y, Wang C, Yang Y, Liu Y, et al. Eatformer: improving vision transformer inspired by evolutionary algorithm. International Journal of Computer Vision. 2024; 132(9):3509-36.
[26]
Gan Y, Li G, Togo R, Maeda K, Ogawa T, Haseyama M. Cross-domain few-shot in-context learning for enhancing traffic sign recognition. In international conference on image processing (ICIP) 2024 (pp. 2564-70). IEEE.
[27]
Chen J, Huang H, Zhang R, Lyu N, Guo Y, Dai HN, et al. Yolo-ts: real-time traffic sign detection with enhanced accuracy using optimized receptive fields and anchor-free fusion. IEEE Transactions on Intelligent Transportation Systems. 2025.
[28]
Azarafza M, Idrees F, Ehteshami BA, Steinmetz C, Henkler S, Rettberg A. Human-in-the-loop reasoning for traffic sign detection: collaborative approach YOLO with video-LLaVA. In future of information and communication conference 2025 (pp. 160-71). Cham: Springer Nature Switzerland.
[29]
Li R, Chen Y, Wang Y, Sun C. YOLO-TSF: a small traffic sign detection algorithm for foggy road scenes. Electronics. 2024; 13(18):1-19.
[30]
Zhang J, Zhang W, Tan C, Li X, Sun Q. Yolo-ppa based efficient traffic sign detection for cruise control in autonomous driving. In proceedings of the international conference on industrial automation and robotics 2024 (pp. 8-13). IEEE.
[31]
Chao B, Liao X, Li G. SEF-YOLO: an improved YOLOv5 algorithm for traffic sign detection. In 43rd Chinese control conference (CCC) 2024 (pp. 6532-6). IEEE.
[32]
Ziqiong L, Wenju L, Liu C, Xiaosong G. A traffic sign recognition method with bi-level routing attention. In 8th international conference on intelligent informatics and biomedical sciences (ICIIBMS) 2023 (pp. 194-9). IEEE.
[33]
Teklesenbet HB, Demoz NH, Jabiro IH, Tesfay YR, Badidi E. Real-time road signs detection and recognition for enhanced road safety. In 15th international conference on innovations in information technology (IIT) 2023 (pp. 132-7). IEEE.
[34]
Yang Z, Yu J. Traffic sign detection and recognition based on an improved YOLOv5. In 4th international conference on big data & artificial intelligence & software engineering (ICBASE) 2023 (pp. 314-8). IEEE.
[35]
Xu J, Huang Y, Ying D. Traffic sign detection and recognition using multi-frame embedding of video-log images. Remote Sensing. 2023; 15(12):1-26.
[36]
Mészáros R, Sergyán S. Lane detection and traffic sign recognition. In IEEE 17th international symposium on applied computational intelligence and informatics (SACI) 2023 (pp. 745-8). IEEE.
[37]
Zhou F, Zu H, Li Y, Song Y, Liao J, Zheng C. Traffic-sign-detection algorithm based on SK-EVC-YOLO. Mathematics. 2023; 11(18):1-12.
[38]
Lin S, Zhang Z, Tao J, Zhang F, Fan X, Lu Q. Traffic sign detection based on lightweight multiscale feature fusion network. Sustainability. 2022; 14(21):1-18.
[39]
Min W, Liu R, He D, Han Q, Wei Q, Wang Q. Traffic sign recognition based on semantic scene understanding and structural traffic sign location. IEEE Transactions on Intelligent Transportation Systems. 2022; 23(9):15794-807.
[40]
Megalingam RK, Thanigundala K, Musani SR, Nidamanuru H, Gadde L. Indian traffic sign detection and recognition using deep learning. International journal of transportation science and technology. 2023; 12(3):683-99.
[41]
Wang Z, Wang J, Li Y, Wang S. Traffic sign recognition with lightweight two-stage model in complex scenes. IEEE Transactions on Intelligent Transportation Systems. 2020; 23(2):1121-31.
[42]
Liu H, Li X, Gong W. Research on detection and recognition of traffic signs based on convolutional neural networks. International Journal of Swarm Intelligence Research (IJSIR). 2022; 13(1):1-9.
[43]
Sristi T, Shilpi K, Radhey S. Detection of traffic sign using CNN. Recent Trends in Parallel Computing. 2023; 12(2):30-8.
[44]
https://pypi.org/project/labelImg/1.8.6/. Accessed 20 August 2025.
[45]
Woo S, Park J, Lee JY, Kweon IS. Cbam: convolutional block attention module. In proceedings of the European conference on computer vision (ECCV) 2018 (pp. 3-19). IEEE.
[46]
Wang CY, Liao HY, Wu YH, Chen PY, Hsieh JW, Yeh IH. CSPNet: a new backbone that can enhance learning capability of CNN. In proceedings of the conference on computer vision and pattern recognition workshops 2020 (pp. 390-1). IEEE.