改进YOLOv5的智慧课堂人脸检测算法

doi:10.3778/j.issn.1002-8331.2304-0352

摘要/Abstract

摘要： 智慧课堂是人工智能领域热门的应用场景。针对课堂场景下摄像头位置较远且偏，图像中目标存在人脸过小和遮挡导致漏检或错检等问题，提出了一种改进YOLOv5的智慧课堂人脸检测算法YOLOv5-SASA。该算法主要包括三个部分，在backbone层沿用了CSPDarknet53网络，通过在最后的空间池化层中使用BasicRFB模块来有效增强网络的特征提取能力；采用NWD损失函数来提高模型对小目标检测的鲁棒性，同时在head层中引入了独立自注意力机制模块SASA，以解决人脸遮挡的问题，并降低模型的参数量；通过降低中间层通道神经元的数量、调节学习率等方式，对改进的YOLOv5网络进行了优化，以避免模型过拟合。实验结果表明，所提出的方法在WiderFace验证集的easy、medium和hard难度下的效果均优于原网络，分别达到了97.5%、96.3%和86.5%的准确率，能够有效提升课堂场景下人脸检测的精度。

关键词: 智慧课堂, 人脸检测, YOLOv5, 独立自注意力机制

Abstract: The intelligent classroom is a popular application scenario in the field of artificial intelligence. This paper proposes a face detection algorithm based on improved YOLOv5, named YOLOv5-SASA, to address the issues of missed or false detection caused by small or occluded faces in images captured by cameras located far away or at an angle. The algorithm consists of three parts. Firstly, the CSPDarknet53 network is utilized in the backbone layer, and the BasicRFB module is used in the final spatial pooling layer to enhance the network’s feature extraction ability. Secondly, the NWD loss function is employed to improve the model’s robustness in detecting small targets. Thirdly, the independent self-attention mechanism module SASA is introduced in the head layer to address the issue of face occlusion and reduce the model’s parameter count. Finally, the improved YOLOv5 network is optimized by reducing the number of neurons in the middle layer channels and adjusting the learning rate to avoid overfitting. Experimental results demonstrate that the proposed method outperforms the original network in the easy, medium, and hard levels of the WiderFace validation set, achieving accuracies of 97.5%, 96.3%, and 86.5%, respectively, which effectively improves the accuracy of face detection in classroom scenarios.

Key words: smart classroom, face detection, YOLOv5, stand-alone self-attentio

钟源, 袁家政, 李鸿天, 刘宏哲, 徐成. 改进YOLOv5的智慧课堂人脸检测算法[J]. 计算机工程与应用, 2024, 60(11): 251-257.

ZHONG Yuan, YUAN Jiazheng, LI Hongtian, LIU Hongzhe, XU Cheng. Intelligent Classroom Face Detection Algorithm with Improved YOLOv5[J]. Computer Engineering and Applications, 2024, 60(11): 251-257.

参考文献

[1] 彭红超, 祝智庭. 面向智慧课堂的灵活深度学习支架设计研究[J]. 中国电化教育, 2023, 435(4): 23-31.
PENG H C, ZHU Z T. A design research on the scaffolds of flexible deep learning for smart classroom[J]. China Educational Technology, 2023, 435(4): 23-31.
[2] 桑高丽, 肖述笛, 赵启军. 联合软阈值去噪和视频数据融合的低质量3维人脸识别[J]. 中国图象图形学报, 2023, 28(5): 1434-1444.
SANG G L, XIAO S D, ZHAO Q J. Soft threshold denoising and video data fusion-relevant low-quality 3D face recognition[J]. Journal of Image and Graphics, 2023, 28(5): 1434-1444.
[3] CANAL F Z, MüLLER T R, MATIAS J C, et al. A survey on facial emotion recognition techniques: a state-of-the-art literature review[J]. Information Sciences, 2022, 582: 593-617.
[4] KUMAR A, KAUR A, KUMAR M. Face detection techniques: a review[J]. Artificial Intelligence Review, 2019, 52: 927-948.
[5] 巢渊, 刘文汇, 唐寒冰, 等. 基于改进YOLO-v4的室内人脸快速检测方法[J]. 计算机工程与应用, 2022, 58(14): 105-113.
CHAO Y, LIU W H, TANG H B, et al. Fast indoor face detection method based on improved YOLO-v4[J]. Computer Engineering and Applications, 2022, 58(14): 105-113.
[6] NSARI M F, LODI K A. A survey of recent trends in two stage object detection methods[J]. Renewable Power for Sustainable Growth, 2021, 723: 669-677.
[7] GIRSHICK R, DONAHUE J, DARRELL T, et al. Rich feature hierarchies for accurate object detection and semantic segmentation[C]//Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, 2014: 580-587.
[8] HE K, GKIOXARI G, DOLLáR P, et al. Mask R-CNN[C]//Proceedings of the IEEE International Conference on Computer Vision, 2017: 2961-2969.
[9] ZHANG K, ZHANG Z, LI Z, et al. Joint face detection and alignment using multitask cascaded convolutional networks[J]. IEEE Signal Processing Letters, 2016, 23(10): 1499-1503.
[10] DAI L, CHEN H, LI Y, et al. TARDet: two-stage anchor-free rotating object detector in aerial images[C]//Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2022: 4267-4275.
[11] GU M, LIU X, FENG J. Classroom face detection algorithm based on improved MTCNN[J]. Signal, Image and Video Processing, 2022, 16(5): 1355-1362.
[12] ZHANG Y F, LI X, WANG F Y, et al. A comprehensive review of one-stage networks for object detection[C]//Proceedings of the IEEE International Conference on Signal Processing, Communications and Computing, 2021: 1-6.
[13] LI C, LI L, JIANG H, et al. YOLOv6: a single-stage object detection framework for industrial applications[J]. arXiv:2209.
02976, 2022.
[14] WANG C Y, BOCHKOVSKIY A, LIAO H Y M. YOLOv7: trainable bag-of-freebies sets new state-of-the-art for real-time object detectors[J]. arXiv:2207.02696, 2022.
[15] DENG J, GUO J, VERVERAS E, et al. RetinaFace: single-shot multi-level face localisation in the wild[C]//Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2020: 5203-5212.
[16] LI J, WANG Y, WANG C, et al. DSFD: dual shot face detector[C]//Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2019: 5060-5069.
[17] QI D, TAN W, YAO Q, et al. YOLO5Face: why reinventing a face detector[C]//Proceedings of the European Conference on Computer Vision, 2023: 228-244.
[18] VASWANI A, SHAZEER N, PARMAR N, et al. Attention is all you need[C]//Proceedings of the 31st International Conference on Neural Information Processing Systems, 2017: 6000-6010.
[19] KITAEV N, KAISER ?, LEVSKAYA A. Reformer: the efficient transformer[J]. arXiv:2001.04451, 2020.
[20] LI K, WANG Y, ZHANG J, et al. Uniformer: unifying convolution and self-attention for visual recognition[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2023, 45(10): 12581-12600.
[21] LIU S, HUANG D. Receptive field block net for accurate and fast object detection[C]//Proceedings of the European Conference on Computer Vision, 2018: 385-400.
[22] WANG J, XU C, YANG W, et al. A normalized Gaussian Wasserstein distance for tiny object detection[J]. arXiv:2110.
13389, 2021.
[23] RAMACHANDRAN P, PARMAR N, VASWANI A, et al. Stand-alone self-attention in vision models[C]//Proceedings of the 33rd International Conference on Neural Information Processing Systems, 2019: 68-80.
[24] YANG S, LUO P, LOY C C, et al. Wider face: a face detection benchmark[C]//Proceedings of the IEEE Conference on Computer Vision and Pattern recognition, 2016: 5525-5533.
[25] 王欣然, 田启川, 张东. 人脸口罩佩戴检测研究综述[J]. 计算机工程与应用, 2022, 58(10): 13-26.
WANG X R, TIAN Q C, ZHANG D. Review of research on face mask wearing detection[J]. Computer Engineering and Applications, 2022, 58(10): 13-26.
[26] GUO J, DENG J, LATTAS A, et al. Sample and computation redistribution for efficient face detection[J]. arXiv:2105.
04714, 2021.
[27] CHEN W, HUANG H, PENG S, et al. YOLO-face: a real-time face detector[J]. The Visual Computer, 2021, 37: 805-813.