西安建筑科技大学学报(自然科学版)

利用热成像相机预测个体热舒适是无干扰温控的一种途径,有助于建筑节能.而人体热舒适范围在不同年龄与性别之间差异较大,现有文献尚缺乏红外热图中年龄与性别差异的研究.为探究利用深度学习从红外热图中自动识别性别与年龄的可行性,本文建立了红外热图和可见光的人脸数据集,对比了ResNet-50、DenseNet-121、DenseNet-201、Inception-V3四种卷积神经网络的效果,实验结果表明:男女红外热图差异明显,用Inception-V3可达到98.7%的识别准确率; 中青年红外热图差异较小,中老年红外热图差异明显,在分三类时,ResNet-50可获得80.0%的年龄识别准确率; 性别与年龄识别准确率均高于现有文献记载.同时,本文研究了红外滤镜和人脸裁剪对准确率的影响,提出了有助于提高识别精度的人脸红外热图数据采集与处理方法.

Thermal camera is a non-invasive method to predict individual thermal comfort which helps saving HVAC energy. While thermal comfort differs a lot between different groups of age and gender, existing literature lack research on the detection of age and gender using thermal cameras for accurate prediction of human thermal comfort in the built environment. To explore the feasibility of using deep learning to automatically recognize gender and age from thermal images, we establish a dataset of thermal images and visible-light images, study the impacts of algorithms(four convolutional neural networks: ResNet-50, DenseNet-121, DenseNet-201, and Inception-V3), thermal image filters, and image cropping on the recognition accuracy, and compare the recognition performances using thermal images and visible-light images. The results show that the gender classification accuracy can reach 98.7% using Inception-V3, meaning that there is a significant difference between male and female thermal images. The highest age classification accuracy(80.0%)is achieved using ResNet-50 when the dataset is divided into three classes—young, middle-aged, and old. It is noticed that there is little difference between young and middle-aged thermal images but a more obvious difference between the middle-aged and the old ones. The achieved accuracies are higher than or comparable to those in the literature. This study demonstrates that convolutional neural network is suitable for gender and age recognition from thermal images.

引言
1 研究方法
2 实验结果
3 结论

图1 摄影角度示意图<br/>Fig.1 Illustration of the camera angles

图1 摄影角度示意图
Fig.1 Illustration of the camera angles

表1 数据集人员统计<br/>Tab.1 Summary of the age and gender in our dataset

表1 数据集人员统计
Tab.1 Summary of the age and gender in our dataset

表2 实验设计方案
Tab.2 Experimental design

图2 卷积神经网络算法对红外热图进行性别与年龄识别的效果对比<br/>Fig.2 Performance of CNN on the age and gender classification of thermal images

图2 卷积神经网络算法对红外热图进行性别与年龄识别的效果对比
Fig.2 Performance of CNN on the age and gender classification of thermal images

图3 性别识别中男性、女性和总体准确率对比<br/>Fig.3 Comparison of male, female, and total accuracy on gender classification

图3 性别识别中男性、女性和总体准确率对比
Fig.3 Comparison of male, female, and total accuracy on gender classification

图4 男女红外热图对比<br/>Fig.4 Comparison of female and male thermal images

图4 男女红外热图对比
Fig.4 Comparison of female and male thermal images

图5 图像处理方法对识别准确率的影响<br/>Fig.5 Influence of image processing on the recognition accuracies

图5 图像处理方法对识别准确率的影响
Fig.5 Influence of image processing on the recognition accuracies

图6 红外热图与可见光图效果对比<br/>Fig.6 Comparison of the performances using thermal images and visible-light images

图6 红外热图与可见光图效果对比
Fig.6 Comparison of the performances using thermal images and visible-light images

表3 实验结果与文献结果对比<br/>Tab.3 Comparison of our results to those in literature

表3 实验结果与文献结果对比
Tab.3 Comparison of our results to those in literature

[1]GRAHAM L T, PARKINSON T, SCHIAVON S. Lessons learned from 20 years of CBE's occupant surveys[J]. Buildings and Cities, 2021, 2(1): 166-184.
[2]于竞宇, 於蓉, 张琦等. 基于机器学习的养老机构室内环境质量满意度评价模型[J]. 西安建筑科技大学学报(自然科学版), 2020, 52(4): 587-593,609.
[3]KIM J, DE DEAR R. Nonlinear relationships between individual IEQ factors and overall workspace satisfaction[J]. Building and Environment, 2012, 49(1): 33-40.
[4]World Green Building Council. Health, wellbeing & productivity in offices: The next chapter for green building[J].The Architrcts' Journal,2014(11):48.
[5]朱颖心. 如何营造健康舒适的建筑热环境——建筑环境与人体舒适及健康关系的探索[J]. 世界建筑, 2021, 3: 42-46.
[6]徐畅, 李念平, 伍志斌, 等. 夏季不同突变热环境下人员热舒适性实验研究[J]. 科学技术与工程, 2020, 20(29): 12097-12103.
[7]兰丽, 连之伟. 改善睡眠热环境可提高睡眠质量[J]. 科学通报, 2020, 65(7): 533-534.
[8]FANG L, WYON D P, CLAUSEN G, et al. Impact of indoor air temperature and humidity in an office on perceived air quality, SBS symptoms and performance[J]. Indoor Air, Supplement, 2004, 14(S): 74-81.
[9]LAN L, WARGOCKI P, WYON D P等. Effects of thermal discomfort in an office on perceived air quality, SBS symptoms, physiological responses, and human performance[J]. Indoor Air, 2011, 21: 376-390.
[10]周翔, 许玲, 谢建彤, 等. 上海地区某高校办公室人员位移及空调器使用行为研究[J]. 建筑科学, 2020, 36(12): 1-7,73.
[11]杨柳, 杨雯, 郑武幸, 等. 风扇对亚热带气候区民居室内热环境影响分析[J]. 西安建筑科技大学学报(自然科学版), 2016, 48(4): 544-550.
[12]林宇凡, 杨柳, 闫海燕, 等. 中国气候与人体热舒适气候适应研究[J]. 西安建筑科技大学学报(自然科学版), 2014, 46(2): 251-255,265.
[13]KIM J, SCHIAVON S, BRAGER G. Personal comfort models-A new paradigm in thermal comfort for occupant-centric environmental control[J]. Building and Environment, 2018, 132: 114-124.
[14]XIE J, LI H, LI C, et al. Review on occupant-centric thermal comfort sensing, predicting, and controlling[J]. Energy and Buildings, 2020, 226: 110392.
[15]李潇婧, 刘一航, 刘朋举, 等. 计算机视觉视频图像处理在暖通空调控制信号采集领域的应用[J]. 暖通空调, 2021, 51(6): 1-12.
[16]YANG B, LI X, HOU Y, et al. Non-invasive(non-contact)measurements of human thermal physiology signals and thermal comfort/discomfort poses: A review[J]. Energy and Buildings, 2020, 224: 110261.
[17]CHENG X, YANG B, OLOFSSON T, et al. A pilot study of online non-invasive measuring technology based on video magnification to determine skin temperature[J]. Building and Environment, 2017, 121: 1-10.
[18]CHENG X, YANG B, HEDMAN A, et al. NIDL: A pilot study of contactless measurement of skin temperature for intelligent building[J]. Energy and Buildings, 2019, 198: 340-352.
[19]YANG B, CHENG X, DAI D, et al. Real-time and contactless measurements of thermal discomfort based on human poses for energy efficient control of buildings[J]. Building and Environment, 2019, 162(March): 106284.
[20]张文利, 郭向, 杨堃等. 面向室内环境控制的人员信息检测系统的设计与实现_张文利[J]. 北京工业大学学报, 2020, 46(5): 456-464.
[21]卢知非,刘浩宇,陈文亮等. 红外人体测温精度补偿方法研究[J]. 红外技术, 2021, 43(9):895-901.
[22]RANJAN J, SCOTT J. Thermal sense: determining dynamic thermal comfort preferences using thermographic imaging[C]//Proceedings of the 2016 ACM International Joint Conference on Pervasive and Ubiquitous Computing.Germany: Heidelberg,2016.
[23]BURZO M, ABOUELENIEN M, ALSTINE D Van, et al. Thermal discomfort detection using thermal imaging[C]//ASME 2017 International Mechanical Engineering Congress and Exposition.USA, Flordia:ASME.2017.
[24]METZMACHER H, WÖLKI D, SCHMIDT C, et al. Real-time human skin temperature analysis using thermal image recognition for thermal comfort assessment[J]. Energy and Buildings, 2018, 158: 1063-1078.
[25]PAVLIN B, PERNIGOTTO G, CAPPELLETTI F, et al. Real-time monitoring of occupants' thermal comfort through infrared imaging: A preliminary study[J]. Buildings, 2017, 7(10): 1-11.
[26]陈庆财, 鹿伟, 张威, 等. 基于人工智能技术预测热感觉的室内热环境控制[J]. 建筑技术, 2019, 50(2): 253-255.
[27]WANG Z, DE DEAR R, LUO M, et al. Individual difference in thermal comfort: A literature review[J]. Building and Environment, 2018, 138: 181-193.
[28]WANG Z, YU H, LUO M, et al. Predicting older people's thermal sensation in building environment through a machine learning approach: Modelling, interpretation, and application[J]. Building and Environment, 2019, 161: 106231.
[29]CHEN C, ROSS A. Evaluation of gender classification methods on thermal and near-infrared face images[C]//2011 International Joint Conference on Biometrics(IJCB).USA, Washington,DC:[s.n.]2011.
[30]WANG S, GAO Z, HE S, et al. Gender recognition from visible and thermal infrared facial images[J]. Multimedia Tools and Applications, 2016, 75: 8419-8442.
[31]NGUYEN D T, KIM K W, HONG H G, et al. Gender recognition from human-body images using visible-light and thermal camera videos based on a convolutional neural network for image feature extraction[J]. Sensors(Switzerland), 2017, 17(3):637-658.
[32]张军挺. 人脸检测及人脸年龄与性别识别方法[D]. 合肥:中国科学技术大学, 2017.
[33]李超. 基于深度学习的人脸性别识别与年龄段估计的研究与实现[D]. 昆明:云南大学, 2019.
[34]HUYNH H T, NGUYEN H. Joint age estimation and gender classification of Asian faces using wide resNet[J]. SN Computer Science, 2020, 1: 284.
[35]刘玉妹. 基于人脸图像的性别分类[D]. 石家庄:河北师范大学, 2019.
[36]张珂. 基于卷积神经网络的人脸检测和人脸属性识别研究[D]. 济南:山东大学, 2019.
[37]魏操. 基于卷积神经网络的图像分类算法研究[D]. 成都:成都理工大学, 2019.
[38]PAN J, YANG Q. A Survey on transfer learning[J]. IEEE Transactions on Knowledge and Data Engineering, 2010, 22(10): 1345-1359.
[39]MEMIS S, ARSLAN B, BATUR O Z, et al. A comparative study of deep learning methods on food classification problem[C]//2020 Innovations in Intelligent Systems and Applications Conference(ASYU).[s.n.],2020.
[40]HE K, ZHANG X, REN S, et al. Deep residual learning for image recognition[C]//2016 IEEE Conference on Computer Vision and Pattern Recognition(CVPR).USA, Las Vegas:IEEE,2016.
[41]HUANG G, LIU Z, VAN DER MAATEN L, et al. Densely connected convolutional networks[C]//30th IEEE Conference on Computer Vision and Pattern Recognition, CVPR 2017,USA,Hawaii:IEEE,2017.
[42]SZEGEDY C, VANHOUCKE V, IOFFE S, et al. Rethinking the Inception Architecture for Computer Vision[C]//Proceedings of the IEEE Computer Society Conference on Computer Vision and Pattern Recognition.USA,Washington,DC:IEEE,2002.

备注

引言

1 研究方法

2 实验结果

3 结论

期刊信息