RPN Prototype Alignment For Domain Adaptive Object Detector

Zhang, Yixin; Wang, Zhong Lin; Mao, Yushi

doi:10.1109/cvpr46437.2021.01224

Cited by 94 publications

(30 citation statements)

References 43 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…To increase discriminability, more recent DA methods attempt to investigate the data structure in unlabeled target domain. Self-training as a typical approach generates target domain pseudo labels [12,26,29,36,39,43,47,49,58,89,92,96,97]. Another category is to construct the protoptypes [4,50,77,78,88,94] or cluster centers [10,24,66] across domains and then perform class-wise alignment.…”

Section: Unsupervised Domain Adaptationmentioning

confidence: 99%

“…Another category is to construct the protoptypes [4,50,77,78,88,94] or cluster centers [10,24,66] across domains and then perform class-wise alignment. Most of these approaches use a fixed probability threshold [12,36,50,58,78], a dynamic probability threshold [47,89], a fixed sample ratio [29,49,92], a dynamic sample ratio [58,96,97], or a threshold of other metrics [4,14,26,39] to choose trustworthy samples (i.e., highconfidence samples) and reject other low-confidence samples. To mitigate the harmful effect of noisy labels, it is reasonable to only utilize the reliable samples.…”

Section: Unsupervised Domain Adaptationmentioning

confidence: 99%

See 1 more Smart Citation

Low-confidence Samples Matter for Domain Adaptation

Zhang¹,

Li²,

Wang³

2022

Preprint

Self Cite

View full text Add to dashboard Cite

Domain adaptation (DA) aims to transfer knowledge from a label-rich source domain to a related but labelscarce target domain. The conventional DA strategy is to align the feature distributions of the two domains. Recently, increasing researches have focused on self-training or other semi-supervised algorithms to explore the data structure of the target domain. However, the bulk of them depend largely on confident samples in order to build reliable pseudo labels, prototypes or cluster centers. Representing the target data structure in such a way would overlook the huge low-confidence samples, resulting in suboptimal transferability that is biased towards the samples similar to the source domain. To overcome this issue, we propose a novel contrastive learning method by processing low-confidence samples, which encourages the model to make use of the target data structure through the instance discrimination process. To be specific, we create positive and negative pairs only using low-confidence samples, and then re-represent the original features with the classifier weights rather than directly utilizing them, which can better encode the task-specific semantic information. Furthermore, we combine cross-domain mixup to augment the proposed contrastive loss. Consequently, the domain gap can be well bridged through contrastive learning of intermediate representations across domains. We evaluate the proposed method in both unsupervised and semi-supervised DA settings, and extensive experimental results on benchmarks reveal that our method is effective and achieves state-ofthe-art performance. The code can be found in https: //github.com/zhyx12/MixLRCo.

show abstract

Section: Unsupervised Domain Adaptationmentioning

confidence: 99%

Section: Unsupervised Domain Adaptationmentioning

confidence: 99%

Low-confidence Samples Matter for Domain Adaptation

Zhang¹,

Li²,

Wang³

2022

Preprint

Self Cite

View full text Add to dashboard Cite

show abstract

“…In this section, we conduct an extra experiment on Synthetic-to-Real scenes to simply verify whether our proposed method is effective for the object detection task. In particular, we choose the recent RPA [76] as the baseline and apply our PCL in the second stage. Table 10 gives the results.…”

Section: Domain Adaptive Object Detectionmentioning

confidence: 99%

Probabilistic Contrastive Learning for Domain Adaptation

Li¹,

Zhang²,

Wang³

et al. 2021

Preprint

Self Cite

View full text Add to dashboard Cite

Recent feature contrastive learning (FCL) has shown promising performance in self-supervised representation learning. For domain adaptation, however, FCL cannot show overwhelming gains since the class weights are not involved during optimization, which does not guarantee the produced features to be clustered around the class weights learned from source data. To tackle this issue, we propose a novel probability contrastive learning (PCL) in this paper, which not only produces compact features but also enforces them to be distributed around the class weights. Specifically, we propose to use the output probabilities after softmax to perform contrastive learning instead of the extracted features and remove the 2 normalization in the traditional FCL. In this way, the probability will approximate the onehot form, thereby narrowing the distance between the features and the class weights. Our proposed PCL is simple and effective. We conduct extensive experiments on two domain adaptation tasks, i.e., unsupervised domain adaptation and semi-supervised domain adaptation. The results on multiple datasets demonstrate that our PCL can consistently get considerable gains and achieves the state-of-theart performance. In addition, our method also obtains considerable gains on semi-supervised tasks when labeled data is scarce.

show abstract

“…Early works [5,15,31] adopt a pixel-to-pixel adaptation in terms of hierarchical features, yielding a global alignment of the whole image with per-pixel adaptation. Some works [5,41,46] focus on foreground objects and conduct more precise adaptation on those regions of interest. Recently, some works [3,36,41,46,47] aim to align the cross-domain class-conditional distribution in the implicit feature space and achieve adaptation in a category-tocategory manner.…”

Section: Introductionmentioning

confidence: 99%

“…Some works [5,41,46] focus on foreground objects and conduct more precise adaptation on those regions of interest. Recently, some works [3,36,41,46,47] aim to align the cross-domain class-conditional distribution in the implicit feature space and achieve adaptation in a category-tocategory manner. These works model category centers with prototypes and minimize the distance of cross-domain prototypes to bridge the domain gap at the category level.…”

Section: Introductionmentioning

confidence: 99%

SIGMA: Semantic-complete Graph Matching for Domain Adaptive Object Detection

Li¹,

Liu²,

Yuan³

2022

Preprint

View full text Add to dashboard Cite

Domain Adaptive Object Detection (DAOD) leverages a labeled domain to learn an object detector generalizing to a novel domain free of annotations. Recent advances align class-conditional distributions by narrowing down crossdomain prototypes (class centers). Though great success, they ignore the significant within-class variance and the domain-mismatched semantics within the training batch, leading to a sub-optimal adaptation. To overcome these challenges, we propose a novel SemantIc-complete Graph MAtching (SIGMA) framework for DAOD, which completes mismatched semantics and reformulates the adaptation with graph matching. Specifically, we design a Graph-embedded Semantic Completion module (GSC) that completes mismatched semantics through generating hallucination graph nodes in missing categories. Then, we establish crossimage graphs to model class-conditional distributions and learn a graph-guided memory bank for better semantic completion in turn. After representing the source and target data as graphs, we reformulate the adaptation as a graph matching problem, i.e., finding well-matched node pairs across graphs to reduce the domain gap, which is solved with a novel Bipartite Graph Matching adaptor (BGM). In a nutshell, we utilize graph nodes to establish semantic-aware node affinity and leverage graph edges as quadratic constraints in a structure-aware matching loss, achieving finegrained adaptation with a node-to-node graph matching. Extensive experiments verify that SIGMA outperforms existing works significantly. Our code is available at https: //github.com/CityU-AIM-Group/SIGMA.

show abstract

RPN Prototype Alignment For Domain Adaptive Object Detector

Cited by 94 publications

References 43 publications

Low-confidence Samples Matter for Domain Adaptation

Low-confidence Samples Matter for Domain Adaptation

Probabilistic Contrastive Learning for Domain Adaptation

SIGMA: Semantic-complete Graph Matching for Domain Adaptive Object Detection

Contact Info

Product

Resources

About