+高级检索
首页 > 优先出版
PDF HTML阅读 XML下载 导出引用 引用提醒
基于VPU加速的嵌入式实时人脸检测系统设计与实现
作者:
作者单位:

1.西塔学院 西南大学;2.信息与通信工程学院 电子科技大学;3.人工智能学院 西南大学

中图分类号:

TP391.4

基金项目:

国家重点研发计划项目(2018YFB1306603)；国家自然科学(61672436)。


Design and implementation of embedded real-time face detection system based on VPU acceleration
Author:
Affiliation:

1.WESTA College,Southwest University;2.School of Information and Communication Engineering,University of Electronic Science and Technology of China;3.College of Artificial Intelligence,Southwest University

  • 摘要
    • | |
  • 访问统计
    • |
  • 参考文献 [30]
    • | | | |
  • 文章评论
    摘要:

    智能设备高昂的设计费用和庞大的计算资源需求成为在便携式、低功耗设备上实现深度学习算法及其应用的主要障碍。本文基于树莓派平台，借助Intel的视频处理器(VPU)低功耗加速模块，设计并实现了基于残差特征提取模块CNN模型的实时人脸检测系统。结果表明，相较于单纯使用树莓派CPU进行计算，本文方法在视频流中检测人脸和检测人脸关键点的实验中分别实现了18.62倍和17.46倍的加速，让在便携式设备中实现快速、实时、在线的人脸检测和特征点提取成为现实，同时为使用便携式、低功耗设备运行深度学习算法提供了一种确实可行的方案。

    Abstract:

    The high design cost and huge computing resource demand of intelligent devices have become the main obstacles to the implementation of deep learning algorithm and its application in portable and low-power devices. In order to solve this dilemma, in this paper, based on the raspberry PI platform and with the help of Intel Video Processing Unit (VPU) low-power acceleration module, a real-time face detection system based on CNN model with residual feature extraction module was designed and implemented. The experimental results showed that, compared with using Central Processing Unit (CPU) of raspberry PI alone, the proposed method achieved 18.62 times and 17.46 times acceleration respectively in the experiments of face detection and face alignment detection in video stream. It realized the fast, real-time and online face detection and face alignment extraction in portable devices. Meanwhile, it also provided a feasible scheme for the operation of deep learning algorithm using portable and low power devices.

    参考文献
    [1] Lecun Y, Bottou L, Bengio Y, et al. Gradient-based learning applied to document recognition [J]. Proceedings of the IEEE, 1998, 86(11): 2278?2324.
    [2] Krizhevsky A, Sutskever I, Hinton G E. Imagenet classification with deep convolu- tional neural networks [J]. Advances in neural information processing systems, 2012, 25(2): 1097?1105.
    [3] Yan Z, Zhang H, Piramuthu R, et al. HD-CNN: hierarchical deep convolutional neural networks for large scale visual recognition [C]//Proceedings of the IEEE international Conference on Computer Vision. Santiago: IEEE, 2015: 2740?2748.
    [4] Gidaris S, Komodakis N. Object detection via a multi-region and semantic segmentation- aware cnn model [C]//Proceedings of the IEEE International Conference on Computer Vision. Santiago: IEEE, 2015: 1134-1142.
    [5] Bappy J H, Roy-Chowdhury A K. CNN based region proposals for efficient object detect [C]//Proionceedings of IEEE International Conference on Image Processing. Phoenix: IEEE, 2016: 3658?3662.
    [6] Lin Y, LI J, Wang H. DCNN-GAN: Recon- structing Realistic Image from fMRI [EB/OL]. (2016-08-26)[2020-03-10].https://arxiv.org/abs/1901.07368.
    [7] 张圣祥, 郑力新, 朱建清, 潘书万. 采用深度学习的快速超分辨率图像重建方法[J]. 华侨大学学报, 2019, 2(40): 245-250ZHANG Shengxiang, ZHENG Lixin, ZHU Jianqing, PAN Shuwan. Fast super-resolution image reconstruction method using deep learning [J].Journal of Huaqiao University, 2019, 2(40): 245-250
    [8] 许晓. 基于深度学习的活体人脸检测算法研究[D]. 北京:北京工业大学, 2016.U Xiao. Face liveness detection algorithm based on deep learning[D]. Beijing: Beijing University of Technology, 2016.
    [9] 王科俊, 赵彦东, 邢向磊．深度学习在无人驾驶汽车领域应用的研究进展[J].智能系统学报, 2018, 13(01):55?69.ANG Kejun, ZHAO Yandong, XING Xianglei. Deep learning in driverless vehicles[J]. CAAI transactions on intelligent systems, 2018, 13(1): 55–69.
    [10] 朱柳依．结合模板匹配与单样本深度学习的货架商品定位与识别技术研究[D].浙江:浙江大学, 2018.ZHU Liuyi. Research on grocery product detection and recognition technology by template matching and one-shot deep learning[D]. Zhejiang: Zhejiang University, 2018.
    [11] 张印, 董兰芳, 王建富. 基于GPU的人脸检测和特征点定位研究[J]. 电子技术, 2014, 9:38?42.HANG Yin, DONG Lanfang, WANG Jianfu. Face detection and feature localization based on GPU. [J]. Electronic Technology, 2014, 9:38?42.
    [12] Ren S, He K, 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]//Proceedings of IEEE Conference on Computer Vision and Pattern Recognition. Las Vegas: IEEE, 2016: 779?788.
    [14] Redmon J, Farhadi A. YOLO9000: better, faster, stronger [C]//Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition. Honolulu: IEEE, 2017: 7263?7271.
    [15] Redmon J, Farhadi A. Yolov3: An incremental improvement [EB/OL]. [2020-03-10]. https://arxiv.org/abs/ 1804.02767.
    [16] Liu W, Anguelov D, Erhan D, et al. SSD: Single shot multibox detector [C]//Proceedings of European Conference on Computer Vision. Amsterdam: Springer, Cham. 2016: 21?37.
    [17] Tan M, Le Q V. EfficientNet: Rethinking Model Scaling for Convolutional Neural Networks[J]. International Conference on Machine Learning, 2019.
    [18] Sun Y, Wang X, Tang X. Deep convolutional network cascade for facial point detection [C]//Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition. Portland: IEEE, 2013: 3476?3483.
    [19] Zhou E, Fan H, Cao Z, et al. Extensive facial landmark localization with coarse-to-fine convolutional network cascade [C]//Proceedings of the IEEE International Conference on Computer Vision Workshops. Sydney: IEEE, 2013: 386?391.
    [20] 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.
    [21] Feng Z H, Kittler J, Awais M, et al. Wing loss for robust facial landmark localisation with convolutional neural networks [C]//Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition. Salt Lake City: IEEE, 2018: 2235?2245.
    [22] 李文胜. 基于树莓派的嵌入式Linux开发教学探索[J]. 电子技术与软件工程, 2014(9): 219?220.I Wensheng. Research on embedded Linux Development teaching based on Raspberry PI [J]. Electronic Technology and Software Engineering, 2014(9): 219?220.
    [23] 罗佳伟, 孙建梅, 徐国旭. 基于树莓派和深度学习技术的智能遥控车的设计与实现[J].计算机产品与流通,2018(01): 185-185.UO Jiawei, SUN Jianmei. XU Guoxu. Design and Implementation of intelligent remote control Vehicle based on Raspberry PI and Deep learning technology [J].Computer Products and Circulation, 2018(01): 185-185.
    [24] 宋凯, 姚嘉明, 李静. 基于树莓派的智能家居控制开关的研究[J]. 电子技术与软件工程, 2015(21): 140?141.ONG Kai, YAO Jiaming, LI Jing. Research on smart Home Control Switch based on Raspberry PI [J]. Electronic Technology and Software Engineering, 2015(21): 140?141.
    [25] 刘华, 田占生, 冯宇飞. 基于树莓派的智能家居语音控制系统[J]. 制造业自动化, 2018, 40(10): 134?137.iu Hua, TIAN Zhansheng, FENG Yufei. Voice Control system for Smart Home Based on Raspberry PI [J]. Manufacturing Automation, 2018, 40(10): 134?137.
    [26] 黄宏敏, 张明森, 詹瑞典. 基于FPGA的YOLO网络的分片加速度方法[J]. 电子世界, 2020(8): 66-67.HUANG Hongmin, ZHANG Minsen, ZHAN Swede. Sharding acceleration method based on FPGA YOLO Network [J]. Electronic World, 2020(8): 66-67.
    [27] 刘勤让, 刘崇阳. 利用参数稀疏性的卷积神经网络计算优化及其FPGA加速器设计[J]. 电子与信息学报, 2018, 40(6): 102?108.IU Qinrang, LIU Chongyang. Calculation optimization for convolutional neural networks and FPGA-based Accelerator design using the parameters sparsity[J]. Journal of Electronics and Information Technology, 2018, 40(6): 102?108.
    [28] 张榜, 来金梅. 一种基于FPGA的卷积神经网络加速器的设计与实现[J]. 复旦学:自然科学版, 2018, 57(2): 236?242.HANG Bang, LAI Jinmei. Design and implementation of a FPGA-based accelerator for convolutional neural networks[J]. Journal of Fudan University (Natural Science). 2018, 57(2):236-242.
    [29] XU J. 斯坦福目标检测深度学习指南[J]. 机器人产业, 2017(06): 20?26.U J. Stanford''s Guide to Deep Learning for Target Detection [J]. Robotics Industry, 2017(06): 20?26.
    [30] He K, Zhang X, Ren S, et al. Deep residual learning for image recognition [C]//Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition. Las Vegas: IEEE, 2016: 770?77.
    相似文献
    引证文献
    网友评论
    网友评论
    分享到微博
    发 布
引用本文

复制
分享
文章指标
  • 点击次数:1405
  • 下载次数: 0
  • HTML阅读次数: 0
  • 引用次数: 0
历史
  • 收稿日期:2020-09-04
  • 最后修改日期:2020-11-03
  • 录用日期:2020-11-16
文章二维码
您是第11637550 位访问者
主管单位：中华人民共和国教育部
主办单位：重庆大学
地址：重庆市沙坪坝正街174号
电话：
重庆大学学报 ® 2025 版权所有