Domain adaptation(DA)for semantic segmen-tation aims to reduce the annotation burden for the dense pixel-level prediction task.It focuses on tackling the domain gap problem and manages to transfer knowledge learned from abundant source data to new target scenes.Although recent works have achieved rapid progress in this field,they still underperform fully supervised models with a large margin due to the absence of any available hints in the target domain.Considering that few-shot labels are cheap to obtain in practical applications,we attempt to leverage them to mitigate the performance gap between DA and fully supervised methods.The key to this problem is to leverage the few-shot labels to learn robust domain-invariant predictions effectively.To this end,we first design a data perturbation strategy to enhance the robustness of the representations.Furthermore,a transferable prototype module is proposed to bridge the domain gap based on the source data and few-shot targets.By means of these proposed methods,our approach can perform on par with the fully supervised models to some extent.We conduct extensive experiments to demonstrate the effectiveness of the proposed methods and report the state-of-the-art performance on two popular DA tasks,i.e.,from GTA5 to Cityscapes and SYNTHIA to Cityscapes.

Junsong FAN;Yuxi WANG;He GUAN;Chunfeng SONG;Zhaoxiang ZHANG

Center for Research on Intelligent Perception and Computing,National Laboratory of Pattern Recognition,Institute of Automation,Chinese Academy of Sciences,Beijing 100190,China;University of Chinese Academy of Sciences,Beijing 100049,China;Centre for Artificial Intelligence and Robotics,HKISI_CAS,HongKong 999077,China

计算机科学前沿

[1]Junsong FAN;Yuxi WANG;He GUAN;Chunfeng SONG;Zhaoxiang ZHANG-.Toward few-shot domain adaptation with perturbation-invariant representation and transferable prototypes)[J].计算机科学前沿,2022(03):81-91

underperform

Toward,few,shot,domain,adaptation,perturbation,invariant,transferable,prototypes,Domain,DA,semantic,segmen,aims,reduce,annotation,burden,dense,pixel,level,It,focuses,tackling,gap,problem,manages,knowledge,learned,from,abundant,source,data,new,scenes,Although,recent,works,have,achieved,rapid,progress,this,field,they,still,fully,supervised,models,large,margin,due,absence,any,available,hints,Considering,that,labels,are,cheap,obtain,practical,applications,attempt,leverage,them,mitigate,performance,between,methods,key,predictions,effectively,end,first,design,strategy,enhance,robustness,representations,Furthermore,module,proposed,bridge,targets,By,means,these,approach,can,par,some,extent,We,conduct,extensive,experiments,demonstrate,effectiveness,report,state,art,two,popular,tasks,GTA5,Cityscapes,SYNTHIA

0.574169