<address id="xrvvb"><listing id="xrvvb"></listing></address>

          <form id="xrvvb"></form>

          <address id="xrvvb"></address>

          華中科技大學學報(自然科學版) 2020, Vol. 48 Issue (9): 70-75 DOI10.13245/j.hust.200912

          欄目:電子與信息工程
          基于自注意力的SSD圖像目標檢測算法
          儲岳中 , 黃 勇 , 張學鋒 , 劉 恒
          安徽工業大學計算機科學與技術學院,安徽 馬鞍山 243002
          摘要 基于深度學習的方法,運用單次多框檢測器(SSD)目標檢測框架和自注意力機制,針對施工人員佩戴安全帽數據集進行神經網絡訓練.通過調整原始SSD目標檢測框架中的參數,并向SSD目標檢測框架中添加自注意力模塊來計算特征圖中像素點之間相互影響,以提高算法對目標檢測的關注度,擴大卷積神經網絡的感受野,從而提高目標檢測的準確率.實驗結果表明:改進算法在應對小目標檢測以及目標之間的遮擋方面有很好的適應性,同時與其他檢測算法相比,檢測成功率有明顯提高.
          關鍵詞 單次多框檢測器(SSD) ;卷積神經網絡 ;自注意力 ;目標檢測 ;安全帽檢測
          SSD image target detection algorithm based on self-attention
          CHU Yuezhong , HUANG Yong , ZHANG Xuefeng , LIU Heng
          School of Computer Science and Technology,Anhui University of Technology,Ma’anshan 243002,Anhui China
          Abstract Aiming at supervising employees to wear safety helmets,a neural network was trained based on deep learning method, the single shot multibox detector (SSD) target detection framework and self-attention mechanism were used to train the neural network.By adjusting the parameters of the original SSD target detection framework,the self-attention module was added to the SSD target detection framework.The mutual influence between pixels in the feature map could be calculated,so as to improve the algorithm's attention to target detection and expand the receptive field of convolutional neural network,and the accuracy of target detection was improved.Experimental results show that the improved algorithm has good adaptability to small target detection and occlusion between targets,and the detection accuracy is significantly improved compared with other detection algorithms.
          Keywords single shot multibox detector (SSD) ; convolutional neural network ; self-attention ; object detection ; hard hat wear detection
          基金資助國家自然科學基金資助項目(61971004);安徽省高校自然科學研究項目(KJ2017ZD05)

          中圖分類號TP391
          文獻標志碼A
          文章編號1671-4512(2020)09-0070-06
          參考文獻
          [1] 黃心漢,蘇豪,彭剛,等.基于卷積神經網絡的目標識別及姿態檢測[J].華中科技大學學報(自然科學版),2017,45(10):7-11.
          [2] RUBAIYAT A H M,TOMA T T,KALANTARI- KHANDANI M,et al.Automatic detection of helmet uses for construction safety[C]// Proc of International Confer- ence on Web Intelligence Workshops.Washington:IEEE Press,2017:135-142.
          [3] 賈峻蘇,鮑慶潔,唐慧明.基于可變形部件模型的安全頭盔佩戴檢測[J].計算機應用研究,2016,33(3):953-956.
          [4] 劉曉慧,葉西寧.膚色檢測和Hu矩在安全帽識別中的應用[J].華東理工大學學報(自然科學版),2014,40(3):365-370.
          [5] DALAL N,TRIGGS B.Histograms of oriented gradients for human detection[C]// Proc of the IEEE Computer Society Conference on Computer Vision and Pattern Recognition.Washington:IEEE,2005:886-893.
          [6] GIRSHICK R,DONAHUE J,DARRELL T,et al.Rich feature hierarchies for accurate object detection and se- mantic segmentation[C]// Proc of the IEEE Conference on Computer Vision and Pattern Recognition.Washington:IEEE,2013:580-587.
          [7] GIRSHICK R.FastR-CNN[C]// Proc of the IEEE Inter- national Conference on Computer Vision.Washington:IEEE,2015:1440-1448.
          [8] REN S,HE K,GIRSHICK R,et al.Faster R-CNN:towards real-time object detection with region proposal networks [C]// Advances in Neural Information Proc- essing Systems.Boston:MIT,2015:91-99.
          [9] REDMON J,DIVVALA S,GIRSHICK R,et al.You only look once:unified,real-time object detection[C]// Proc of IEEE Conference on Computer Vision and Pattern Recognition.Washington:IEEE,2016:779-788.
          [10] LIU W,ANGUELOV D,ERHAN D,et al.SSD:single shot multibox detector[C]// Proc of the 14th European Conference on Computer Vision.New York:Springer,2016:21-37.
          [11] ZHANG H,IAN G,METAXAS D,et al.Self-attention generative adversarial networks[C]// Proceedings of the 36th International Conference on Machine Learning.New York:ACM,2019:7354-7363.
          [12] ZHU L,CHEN Y,GHAMISI P,et al.Generative adversarial networks for hyperspectral image classifi- cation[J].IEEE Transactions on Geoscience and Remote Sensing,2018,56(9):5046-5063.
          [13] HE K M,ZHANG X Y,REN S Q,et al.Deep residual learning for image recognition[C]// Proc of the IEEE Conference on Computer Vision and Pattern Recognition.Washington:IEEE,2016:770-778.
          [14] EVERINGHAM M,GOOL L V,WILLIAMS C K I,et al.The pascal,visual object classes (VOC) challenge [J].International Journal of Computer Vision,2010,88(2):303-338.
          [15] XIA Y Z,ZHANG BAILING,FRANS COENEN.Face occlusion detection using deep convolutional neural networks[J].International Journal of Pattern Recognition & Artificial Intelligence,2016,30(9):401-408.
          [16] FELZENSZWALB P F,GIRSHICK R B,MCALLESTER D,et al.Object detection with discriminatively trained part-based models[J].IEEE Trans on Pattern Analysis & Machine Intelligence,2010,32(9):1627-1635.
          [17] REDMON J,FARHADI A.YOLO9000:better,faster,stronger[C]// Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition.Washington:IEEE,2017:7263-7271.
          文獻來源
          儲岳中, 黃 勇, 張學鋒, 劉 恒. 基于自注意力的SSD圖像目標檢測算法[J]. 華中科技大學學報(自然科學版), 2020, 48(9): 70-75
          DOI:10.13245/j.hust.200912
          澳门盘口