Deep learning-based face detection method under low illumination conditions in coal mines
-
摘要: 煤矿井下光线昏暗、人工光源干扰等造成监控系统采集到的人脸图像对比度低、人脸特征模糊,传统人脸检测算法应用于煤矿井下时会出现误检、漏检。针对上述问题,提出了一种基于深度学习的煤矿井下低光照人脸检测方法。采用基于无监督学习的生成对抗网络(GAN)对煤矿井下低光照图像进行对比度增强,使用自调整注意力引导的U−Net作为生成器,利用双判别器对全局和局部信息进行引导,并使用自特征保留损失函数来指导训练过程和维护图像中人脸的纹理结构,强化人脸特征,避免出现曝光、人脸细节信息丢失等现象,得到较为清晰的人脸图像;利用RetinaFace人脸检测框架对增强后的人脸特征进行检测,其采用特征金字塔结构和单阶段检测模式对人脸图像进行检测,在基本不增加计算量的同时,提高对小尺度人脸检测的能力。在公开低光照人脸数据集DARK FACE和自建煤矿井下人脸数据集上的实验结果表明,该方法提高了图像对比度,清晰地恢复了图像中的人脸特征,在准确率、召回率、平均精度方面均表现较好,有效提高了煤矿井下人脸检测精度。Abstract: The low contrast and blurry facial features of facial images collected by the monitoring system are caused by dim illumination and interference from artificial illumination sources in coal mines. Traditional facial detection algorithms may cause false or missed detections when applied in coal mines. In order to solve the above problems, a deep learning-based face detection method under low illumination conditions in coal mines is proposed. A generative adversarial network (GAN) based on unsupervised learning is used to enhance the contrast of low illumination images in coal mines. A self-adjusting attention guided U-Net is used as the generator, and dual discriminators are used to guide global and local information. The self-feature retention loss function is used to guide the training process and maintain the texture structure of the face in the image and strengthen facial features. It can avoid phenomena such as exposure and loss of facial detail information, and obtain clearer facial images. The RetinaFace face detection framework is used to detect the enhanced facial features. It uses a feature pyramid structure and a single stage detection mode to detect facial images. It improves the capability to detect small-scale faces without increasing computational complexity. The experimental results on the public low illumination face dataset DARK FACE and the self built coal mine underground face dataset show that this method improves image contrast, clearly restores facial features in the image, and performs well in accuracy, recall, and average accuracy, effectively improving the accuracy of coal mine underground face detection.
-
表 1 不同方法下客观评价结果
Table 1. Objective evaluation results of different methods
% 方法 准确率 召回率 平均精度 RetinaFace 91.2 59.4 54.5 本文方法 91.8 64.7 59.0 -
[1] 钱鸣高,许家林,王家臣. 再论煤炭的科学开采[J]. 煤炭学报,2018,43(1):1-13. doi: 10.13225/j.cnki.jccs.2017.4400QIAN Minggao,XU Jialin,WANG Jiachen. Further on the sustainable mining of coal[J]. Journal of China Coal Society,2018,43(1):1-13. doi: 10.13225/j.cnki.jccs.2017.4400 [2] GIRSHICK R,DONAHUE J,DARRELL T,et al. Rich feature hierarchies for accurate object detection and semantic segmentation[C]. IEEE Conference on Computer Vision and Pattern Recognition,Columbus,2014:580-587. [3] 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. doi: 10.1109/TPAMI.2016.2577031 [4] 郑道能. 一种改进的tiny YOLO v3煤矸石快速识别模型[J]. 工矿自动化,2023,49(4):113-119.ZHENG Daoneng. An improved tiny YOLO v3 rapid recognition model for coal-gangue[J]. Journal of Mine Automation,2023,49(4):113-119. [5] BERG A C,FU Chengyang,SZEGEDY C,et al. SSD:single shot multibox detector[C]. European Conference on Computer Vision,Amsterdam,2016:21-37. [6] LI Jian,WANG Yabiao,WANG Changan,et al. DSFD:dual shot face detector[C]. IEEE/CVF Conference on Computer Vision and Pattern Recognition,Los Angeles,2019:5060-5069. [7] TANG Xu,DU D K,HE Zeqiang,et al. PyramidBox:a context-assisted single shot face detector[C]. European Conference on Computer Vision,Munich,2018:797-813. [8] ZHANG Kaipeng,ZHANG Zhanpeng,LI Zhifeng,et al. Joint face detection and alignment using multi-task cascaded convolutional networks[J]. IEEE Signal Processing Letters,2016,23(10):1499-1503. doi: 10.1109/LSP.2016.2603342 [9] DENG Jiankang,GUO Jia,VERVERAS E,et al. Retinaface:single-shot multi-level face localisation in the wild[C]. IEEE/CVF Conference on Computer Vision and Pattern Recognition,Seattle,2020:5203-5212. [10] 刘淇缘. 单阶段复杂人脸检测方法研究[D]. 北京:中国人民公安大学,2021.LIU Qiyuan. Research on one-stage complex face detection methods[D]. Beijing:People's Public Security University of China,2021. [11] YANG Shuo,LUO Ping,LOY C C,et al. Wider face:a face detection benchmark[C]. IEEE Conference on Computer Vision and Pattern Recognition,Las Vegas,2016:5525-5533. [12] 万俊霞,林珊玲,梅婷,等. 基于图像分割和动态直方图均衡的电润湿显示器图像增强算法[J]. 光子学报,2022,51(2):240-250.WAN Junxia,LIN Shanling,MEI Ting,et al. Image enhancement algorithm of electrowetting display based on image segmentation and dynamic histogram equalization[J]. Acta Photonica Sinica,2022,51(2):240-250. [13] REN Xutong,YANG Wenhan,CHENG Wenhuang,et al. LR3M:robust low-light enhancement via low-rank regularized Retinex model[J]. IEEE Transactions on Image Processing,2020,29:5862-5876. doi: 10.1109/TIP.2020.2984098 [14] WANG Lei,FU Guangtao,JIANG Zhuqiang,et al. Low-light image enhancement with attention and multi-level feature fusion[C]. IEEE International Conference on Multimedia & Expo Workshops,Shanghai,2019:276-281. [15] LI Jinjiang,FENG Xiaomei,HUA Zhen. Low-light image enhancement via progressive-recursive network[J]. IEEE Transactions on Circuits and Systems for Video Technology,2021,31(11):4227-4240. doi: 10.1109/TCSVT.2021.3049940 [16] GUO Chunle,LI Chongyi,GUO Jichang,et al. Zero-reference deep curve estimation for low-light image enhancement[C]. IEEE/CVF Conference on Computer Vision and Pattern Recognition,Seattle,2020:1780-1789. [17] LEE H,SOHN K,MIN Dongbo. Unsupervised low-light image enhancement using bright channel prior[J]. IEEE Signal Processing Letters,2020,27:251-255. doi: 10.1109/LSP.2020.2965824 [18] JIANG Yifan,GONG Xinyu,LIU Ding,et al. EnlightenGAN:deep light enhancement without paired supervision[J]. IEEE Transactions on Image Processing,2021,30:2340-2349. doi: 10.1109/TIP.2021.3051462 [19] 刘丹英,刘晓燕. 基于U−Net卷积神经网络的多尺度遥感图像分割算法[J]. 现代电子技术,2023,46(21):44-47. doi: 10.16652/j.issn.1004-373x.2023.21.009LIU Danying,LIU Xiaoyan. Multi-scale remote sensing image segmentation algorithm based on U-net convolutional neural network[J]. Modern Electronics Technique,2023,46(21):44-47. doi: 10.16652/j.issn.1004-373x.2023.21.009 [20] 王照乾,孔韦韦,滕金保,等. DenseNet生成对抗网络低照度图像增强方法[J]. 计算机工程与应用,2022,58(8):214-220.WANG Zhaoqian,KONG Weiwei,TENG Jinbao,et al. Low illumination image enhancement method based on DenseNet GAN[J]. Computer Engineering and Applications,2022,58(8):214-220. [21] DONG Xuan,WANG Guan,PANG Yi,et al. Fast efficient algorithm for enhancement of low lighting video[C]. IEEE International Conference on Multimedia and Expo,Barcelona,2011:1-6. [22] MA Long,MA Tengyu,LIU Risheng,et al. Toward fast,flexible,and robust low-light image enhancement[C]. IEEE/CVF Conference on Computer Vision and Pattern Recognition,New Orleans,2022:5637-5646. -