典型文献
A similarity-guided segmentation model for garbage detection under road scene
文献摘要：
The development of computer vision technology provides a possible path for realizing intelligent control of road sweepers to reduce energy waste in urban street cleaning work.For garbage segmentation of seven categories under road scene,we introduce an efficient deep-learning-based method.Our model follows a lightweight structure with a feature pyramid attention(FPA)module employed in the decoder to enhance feature integration at multi-levels.Besides,a similarity guidance(SG)module is added to the decoder branches,which calculates the cosine distance between learned prototypes and feature maps to guide the segmentation results from a metric learning perspective.Our model has less than 3 M parameters and can run at over 65 FPS in an RTX 2070 GPU.Experimental results demonstrate that our method can yield competitive results in terms of speed and accuracy trade-off,with overall mean intersection-over-union(mIoU)reaching 0.87 and 0.67,respectively,on two garbage data sets we built.Besides,our model can perform acceptable category-balanced segmentation from less than 20 annotated images per category by introducing the SG module.
文献关键词：
中图分类号：
作者姓名：
Caiyun Zheng;Danhua Cao;Cheng Hu
作者机构：
School of Optical and Electronic Information,Huazhong University of Science and Technology,Wuhan 430074,China
文献出处：
引用格式：
[1]Caiyun Zheng;Danhua Cao;Cheng Hu-.A similarity-guided segmentation model for garbage detection under road scene)[J].光电子前沿（英文版）,2022(02):79-95
A类：
sweepers
B类：
similarity,guided,segmentation,model,garbage,detection,under,road,scene,development,computer,vision,technology,provides,possible,path,realizing,intelligent,control,reduce,energy,waste,urban,street,cleaning,work,For,seven,categories,introduce,efficient,deep,learning,method,Our,follows,lightweight,structure,feature,pyramid,attention,FPA,module,employed,decoder,enhance,integration,multi,levels,Besides,guidance,SG,added,branches,which,calculates,cosine,distance,between,learned,prototypes,maps,results,from,metric,perspective,has,less,than,parameters,can,run,FPS,RTX,GPU,Experimental,demonstrate,that,our,yield,competitive,terms,speed,accuracy,trade,off,overall,mean,intersection,union,mIoU,reaching,respectively,two,data,sets,built,perform,acceptable,category,balanced,annotated,images,by,introducing
AB值：
0.656002
相似文献
机标中图分类号，由域田数据科技根据网络公开资料自动分析生成，仅供学习研究参考。
