基于YOLOv8-DF的轻量化驾驶员面部目标检测算法

doi:10.16182/j.issn1004731x.joss.24-0320

摘要/Abstract

摘要：

针对驾驶环境下YOLOv8n检测算法存在计算量、参数量较大的问题，提出一种结构轻量化的驾驶员面部目标检测算法YOLOv8-DF。提出一种轻量级多尺度卷积模块LMCM(lightweight multi-scale convolution module)替换网络中的Conv模块，双通道设计可以降低算法的计算量和参数量，多尺度设计可以丰富网络内部的特征信息；引入轻量级卷积GhostConv、Fasterblock模块与C2f模块融合；提出一种双通道轻量级卷积模块DLCM(dual-channel lightweight convolution module)与SPPF模块融合。实验结果表明：YOLOv8n-DF算法在测试集的平均精确率达到了99.2%，相较YOLOv8n检测算法参数量下降了36.7%，计算量下降了30.9%，帧率达到了282 帧/s。

关键词: 驾驶员面部检测, 轻量级卷积, 多尺度特征融合, 实时性, 目标检测

Abstract:

The YOLOv8n detection algorithm has a large amount of computation and parameters in the driving environment. To address this issue, a lightweight driver facial object detection algorithm YOLOv8-DF was proposed. A lightweight multi-scale convolution module (LMCM) was proposed to replace the Conv module in the network, and the dual-channel design could reduce the computation and parameter quantity of the algorithm; the multi-scale design could enrich the feature information inside the network. The lightweight convolutional GhostConv, Fasterblock module, and C2f module were fused, and a dual-channel lightweight convolution module (DLCM) was fused with the SPPF module. The experimental results show that the average precision of the YOLOv8n-DF algorithm in the test set reaches 99.2%, which is 36.7% lower than that of the YOLOv8n detection algorithm; the amount of computation is reduced by 30.9%, and the frame rate reaches 282 frames/s.

Key words: driver face detection, lightweight convolution, multi-scale feature fusion, real-time performance, object detection

中图分类号:

TP391.4

李明煜,林家泉 . 基于YOLOv8-DF的轻量化驾驶员面部目标检测算法[J]. 系统仿真学报, 2025, 37(8): 2103-2114.

Li Mingyu,Lin Jiaquan . Lightweight Driver Face Object Detection Algorithm Based on YOLOv8-DF[J]. Journal of System Simulation, 2025, 37(8): 2103-2114.

图/表 10

图1

图2

图3

图4

图5

表1

图6

表2

表3

图7

参考文献 34

[1]	Amodio Alessandro, Ermidoro Michele, Maggi D, et al. Automatic Detection of Driver Impairment Based on Pupillary Light Reflex[J]. IEEE Transactions on Intelligent Transportation Systems, 2019, 20(8): 3038-3048.
[2]	邱轶辉, 江琼, 魏玲玲, 等. 基于迁移学习的跨被试脑电疲劳驾驶检测[J]. 南昌大学学报(理科版), 2023, 47(4): 397-402.
	Qiu Yihui, Jiang Qiong, Wei Lingling, et al. Transfer Learning-based Cross-subject EEG Fatigue Driving Detection[J]. Journal of Nanchang University(Natural Science), 2023, 47(4): 397-402.
[3]	Guo Zizheng, Wang Guanglei, Zhou Ming, et al. Monitoring and Detection of Driver Fatigue from Monocular Cameras Based on Yolov5[C]//2022 6th CAA International Conference on Vehicular Control and Intelligence (CVCI). Piscataway: IEEE, 2022: 1-6.
[4]	Sikander Gulbadan, Anwar Shahzad. Driver Fatigue Detection Systems: A Review[J]. IEEE Transactions on Intelligent Transportation Systems, 2019, 20(6): 2339-2352.
[5]	李小平, 白超. 一种基于多模态信息融合的火车司机疲劳驾驶检测方法[J]. 铁道学报, 2022, 44(6): 56-65.
	Li Xiaoping, Bai Chao. A Train Driver Fatigue Driving Detection Method Based on Multi-modal Information Fusion[J]. Journal of the China Railway Society, 2022, 44(6): 56-65.
[6]	Krizhevsky A, Sutskever I, Hinton G E. ImageNet Classification with Deep Convolutional Neural Networks[J]. Communications of the ACM, 2017, 60(6): 84-90.
[7]	Viola P, Jones M. Rapid Object Detection Using a Boosted Cascade of Simple Features[C]//Proceedings of the 2001 IEEE Computer Society Conference on Computer Vision and Pattern Recognition. CVPR 2001. Piscataway: IEEE, 2001: I-511-I-518.
[8]	Dalal N, Triggs B. Histograms of Oriented Gradients for Human Detection[C]//2005 IEEE Computer Society Conference on Computer Vision and Pattern Recognition (CVPR'05). Piscataway: IEEE, 2005: 886-893.
[9]	Felzenszwalb P F, Girshick R B, McAllester D, et al. Object Detection with Discriminatively Trained Part-based Models[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2010, 32(9): 1627-1645.
[10]	Agrawal P, Girshick R, Malik J. Analyzing the Performance of Multilayer Neural Networks for Object Recognition[C]//Computer Vision-ECCV 2014. Cham: Springer International Publishing, 2014: 329-344.
[11]	Girshick R. Fast R-CNN[C]//2015 IEEE International Conference on Computer Vision (ICCV). Piscataway: IEEE, 2015: 1440-1448.
[12]	Ren Shaoqing, He Kaiming, Girshick R, et al. Faster R-CNN: Towards Real-time Object Detection with Region Proposal Networks[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2017, 39(6): 1137-1149.
[13]	Redmon J, Divvala S, Girshick R, et al. You Only Look Once: Unified, Real-time Object Detection[C]//2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR). Piscataway: IEEE, 2016: 779-788.
[14]	Liu Wei, Anguelov D, Erhan D, et al. SSD: Single Shot MultiBox Detector[C]//Computer Vision-ECCV 2016. Cham: Springer International Publishing, 2016: 21-37.
[15]	Carion N, Massa F, Synnaeve G, et al. End-to-end Object Detection with Transformers[C]//Computer Vision-ECCV 2020. Cham: Springer International Publishing, 2020: 213-229.
[16]	Li Kening, Gong Yunbo, Ren Ziliang. A Fatigue Driving Detection Algorithm Based on Facial Multi-feature Fusion[J]. IEEE Access, 2020, 8: 101244-101259.
[17]	Niu Chengwen, Song Yunsheng, Zhao Xinyue. SE-lightweight YOLO: Higher Accuracy in YOLO Detection for Vehicle Inspection[J]. Applied Sciences, 2023, 13(24): 13052.
[18]	苏彤, 王颖, 邓启扬, 等. 基于YOLOv5改进的雾天行人与车辆检测算法[J]. 系统仿真学报, 2024, 36(10): 2413-2422.
	Su Tong, Wang Ying, Deng Qiyang, et al. Improved Foggy Pedestrian and Vehicle Detection Algorithm Based on YOLOv5[J]. Journal of System Simulation, 2024, 36(10): 2413-2422.
[19]	王银, 王飞翔, 孙前来. 多尺度特征融合车辆检测方法[J]. 系统仿真学报, 2022, 34(6): 1219-1229.
	Wang Yin, Wang Feixiang, Sun Qianlai. Vehicle Detection Method Based on Multi Scale Feature Fusion[J]. Journal of System Simulation, 2022, 34(6): 1219-1229.
[20]	Chen Leiyu, Li Shaobo, Bai Qiang, et al. Review of Image Classification Algorithms Based on Convolutional Neural Networks[J]. Remote Sensing, 2021, 13(22): 4712.
[21]	Sun Zhaoyun, Zhai Junzhi, Pei Lili, et al. Automatic Pavement Crack Detection Transformer Based on Convolutional and Sequential Feature Fusion[J]. Sensors, 2023, 23(7): 3772.
[22]	付锦燚, 张自嘉, 孙伟, 等. 改进YOLOv8的航拍图像小目标检测算法[J]. 计算机工程与应用, 2024, 60(6): 100-109.
	Fu Jinyi, Zhang Zijia, Sun Wei, et al. Improved YOLOv8 Small Target Detection Algorithm in Aerial Images[J]. Computer Engineering and Applications, 2024, 60(6): 100-109.
[23]	郭业才, 孙京东, Saha Amitave. 基于YOLOv5改进的航拍图像目标检测算法[J]. 系统仿真学报, 2025, 37(2): 551-562.
	Guo Yecai, Sun Jingdong, Saha Amitave. Improved Target Detection Algorithm for Aerial Images Based on YOLOv5[J]. Journal of System Simulation, 2025, 37(2): 551-562.
[24]	Chen Long, Ding Qiwei, Zou Qin, et al. DenseLightNet: A Light-weight Vehicle Detection Network for Autonomous Driving[J]. IEEE Transactions on Industrial Electronics, 2020, 67(12): 10600-10609.
[25]	Chen Jierun, Kao S H, He Hao, et al. Run, Don't Walk: Chasing Higher FLOPS for Faster Neural Networks[C]//2023 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR). Piscataway: IEEE, 2023: 12021-12031.
[26]	Wang Jian, Zhang Fei, Zhang Yuesong, et al. Lightweight Object Detection Algorithm for UAV Aerial Imagery[J]. Sensors, 2023, 23(13): 5786.
[27]	Han Kai, Wang Yunhe, Tian Qi, et al. GhostNet: More Features from Cheap Operations[C]//2020 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR). Piscataway: IEEE, 2020: 1577-1586.
[28]	Alansari Mohamad, Oussama Abdul Hay, Javed Sajid, et al. GhostFaceNets: Lightweight Face Recognition Model from Cheap Operations[J]. IEEE Access, 2023, 11: 35429-35446.
[29]	Zhang Zicheng, Xu Hongke, Lin Shan. Quantizing YOLOv5 for Real-time Vehicle Detection[J]. IEEE Access, 2023, 11: 145601-145611.
[30]	许晓阳, 高重阳. 改进YOLOv7-tiny的轻量级红外车辆目标检测算法[J]. 计算机工程与应用, 2024, 60(1): 74-83.
	Xu Xiaoyang, Gao Chongyang. Improved YOLOv7-tiny Lightweight Infrared Vehicle Target Detection Algorithm[J]. Computer Engineering and Applications, 2024, 60(1): 74-83.
[31]	Lin T, Maire M, Belongie S, et al. Microsoft COCO: Common Objects in Context[C]//Computer Vision-ECCV 2014. Cham: Springer International Publishing, 2014: 740-755.
[32]	Zhao Yian, Wenyu Lü, Xu Shangliang, et al. DETRs Beat YOLOs on Real-time Object Detection[C]//2024 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR). Piscataway: IEEE, 2024: 16965-16974.
[33]	Li Hulin, Li Jun, Wei Hanbing, et al. Slim-neck by GSConv: A Better Design Paradigm of Detector Architectures for Autonomous Vehicles[EB/OL]. (2022-06-06) [2024-03-24]. .
[34]	Ma Ningning, Zhang Xiangyu, Zheng Haitao, et al. ShuffleNet V2: Practical Guidelines for Efficient CNN Architecture Design[C]//Computer Vision-ECCV 2018. Cham: Springer International Publishing, 2018: 122-138.

LMCM	C2f-FG	SPPF-DLCM	AP0.5/%	参数量/10⁶	计算量/10⁹	权重/MB
×	×	×	99.1	3.006	8.1	6.1
√	×	×	99.2	2.724	7.6	5.5
√	√	×	99.2	2.127	6.5	4.0
√	√	√	99.2	1.897	5.6	3.9

算法	AP0.5/%	AP0.5:0.95/%	准确率/%	召回率/%	参数量/10⁶	计算量/10⁹
YOLOv8n-HGNetV2	99.2	63.8	97.9	97.9	2.5	7.7
YOLOv8n-Slimneck	99.2	64.6	97.3	98.7	2.9	8.0
YOLOv8n-ShuffleNetV2	99.1	62.4	96.6	97.8	1.8	5.1
YOLOv8-DF	99.2	63.7	97.7	98.4	1.9	5.6

算法	AP0.5/%	帧率/(帧/s)	参数量/10⁶	计算量/10⁹	权重/MB
Faster-RCNN	82.6	12	137.1	370.2	110.9
YOLOv3	99.2	29	61.5	154.6	123.5
YOLOv5s	99.1	147	7.0	15.8	14.4
YOLOv7-tiny	98.6	161	6.0	13.2	12.3
YOLOv8n	99.1	284	3.0	8.1	6.1
YOLOv8s	99.1	167	11.1	28.4	22.0
YOLOv8-DF	99.2	282	1.9	5.6	3.9