Uncertainty-Aware Unsupervised Domain Adaptation in Object Detection

Guan, Dayan; Huang, Jiaxing; Xiao, Aoran; Lu, Shijian; Cao, Yanan

doi:10.1109/tmm.2021.3082687

Cited by 90 publications

(30 citation statements)

References 75 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…As discussed earlier, a variety of methods [56], [57], [58], [59], [60], [61], [62], [63], [64], [65], [66], [67], [68], [69], [70], [71], [72], [73], [74], [75], [76], [77], [78], [79], [80] have been proposed for the task of domain adaptive object detection. After a meticulous review of these approaches, we categorize them into the following classes: 1) Adversarial feature learning, 2) Image-to-image translation, 3) Domain randomization, 4) Pseudo-label self-training, 5) Mean-teacher training, and 6) Graph reasoning.…”

Section: Methodsmentioning

confidence: 99%

“…There are several subsequent works such as [66], [76], [63], [54], [96], [70], [100], [106] that utilize the adversarial feature learning strategies discussed in this section. Most of these approaches address cross-domain detection for the application of autonomous driving/surveillance.…”

Section: Additional Methodsmentioning

confidence: 99%

“…This poses unique challenges while addressing unsupervised domain adaptation for object detection models. This has sparked an interest in addressing unsupervised domain adaptation for object detection with many novel approaches proposed very recently [56], [57], [58], [59], [60], [61], [62], [63], [64], [65], [66], [67], [68], [69], [70], [71], [72], [73], [74], [75], [76], [77], [78], [79], [80]. Some of the landmark papers proposed over the years are shown in Fig.…”

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Unsupervised Domain Adaptation of Object Detectors: A Survey

Oza¹,

Sindagi²,

VS³

et al. 2021

Preprint

View full text Add to dashboard Cite

Recent advances in deep learning have led to the development of accurate and efficient models for various computer vision applications such as object classification, semantic segmentation, and object detection. However, learning highly accurate models relies on the availability of datasets with large number of annotated images. Due to this, model performance drops drastically when evaluated on label-scarce datasets having visually distinct images. This issue is commonly referred to as covariate shift or dataset bias. Domain adaptation attempts to address this problem by leveraging domain shift characteristics from labeled data in a related domain when learning a classifier for label-scarce target dataset. There are a plethora of works to adapt object classification and semantic segmentation models to label-scarce target dataset through unsupervised domain adaptation. Considering that object detection is a fundamental task in computer vision, many recent works have recently focused on addressing the domain adaptation issue for object detection as well. In this paper, we provide a brief introduction to the domain adaptation problem for object detection and present an overview of various methods proposed till date for addressing this problem. Furthermore, we highlight strategies proposed for this problem and the associated shortcomings. Subsequently, we identify multiple aspects of the unsupervised domain adaptive detection problem that are most promising for the future research in the area. We believe that this survey shall be valuable to the pattern recognition experts working in the fields of computer vision, biometrics, medical imaging, and autonomous navigation by introducing them to the problem, getting them familiar with the current status of the progress and providing with promising direction for future research.

show abstract

Section: Methodsmentioning

confidence: 99%

Section: Additional Methodsmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Unsupervised Domain Adaptation of Object Detectors: A Survey

Oza¹,

Sindagi²,

VS³

et al. 2021

Preprint

View full text Add to dashboard Cite

show abstract

“…Most existing works can be broadly classified into three categories. The first category is adversarial training based which employs a discriminator to align source and target domains in the feature, output or latent space [26,50,84,52,81,11,99,68,71,86,82,43,20,97]. The second category is image translation based which adapt image appearance to mitigate domain gaps [25,75,12,45,98,27,94,31,29].…”

Section: Related Workmentioning

confidence: 99%

RDA: Robust Domain Adaptation via Fourier Adversarial Attacking

Huang

Guan

Xiao

et al. 2021

Preprint

Self Cite

View full text Add to dashboard Cite

Unsupervised domain adaptation (UDA) involves a supervised loss in a labeled source domain and an unsupervised loss in an unlabeled target domain, which often faces more severe overfitting (than classical supervised learning) as the supervised source loss has clear domain gap and the unsupervised target loss is often noisy due to the lack of annotations. This paper presents RDA, a robust domain adaptation technique that introduces adversarial attacking to mitigate overfitting in UDA. We achieve robust domain adaptation by a novel Fourier adversarial attacking (FAA) method that allows large magnitude of perturbation noises but has minimal modification of image semantics, the former is critical to the effectiveness of its generated adversarial samples due to the existence of 'domain gaps'. Specifically, FAA decomposes images into multiple frequency components (FCs) and generates adversarial samples by just perturbating certain FCs that capture little semantic information. With FAA-generated samples, the training can continue the 'random walk' and drift into an area with a flat loss landscape, leading to more robust domain adaptation. Extensive experiments over multiple domain adaptation tasks show that RDA can work with different computer vision tasks with superior performance.

show abstract

“…To this end, researchers have designed different unsupervised losses on target data for learning a well-performed model in target domain [27,34,35,68,72,75,47,77,5,8]. The existing unsupervised losses can be broadly classified into three categories: 1) adversarial loss that enforces source-like target representations in the feature, output or latent space [27,45,75,47,72,9,91,61,64,77,73,40,33,17,87,30]; 2) image translation loss that translates source images to have target-like styles and appearance [26,68,10,41,90,28,84,32]; and 3) self-training loss that re-trains networks iteratively with confidently pseudo-labelled target samples [96,95,66,92,41,16,31].…”

Section: Introductionmentioning

confidence: 99%

Category Contrast for Unsupervised Domain Adaptation in Visual Tasks

Huang¹,

Guan²,

Xiao³

et al. 2021

Preprint

Self Cite

View full text Add to dashboard Cite

Instance contrast for unsupervised representation learning has achieved great success in recent years. In this work, we explore the idea of instance contrastive learning in unsupervised domain adaptation (UDA) and propose a novel Category Contrast technique (CaCo) that introduces semantic priors on top of instance discrimination for visual UDA tasks. By considering instance contrastive learning as a dictionary look-up operation, we construct a semantics-aware dictionary with samples from both source and target domains where each target sample is assigned a (pseudo) category label based on the category priors of source samples. This allows category contrastive learning (between target queries and the category-level dictionary) for category-discriminative yet domain-invariant feature representations: samples of the same category (from either source or target domain) are pulled closer while those of different categories are pushed apart simultaneously. Extensive UDA experiments in multiple visual tasks (e.g., segmentation, classification and detection) show that the simple implementation of CaCo achieves superior performance as compared with the highly-optimized state-of-the-art methods. Analytically and empirically, the experiments also demonstrate that CaCo is complementary to existing UDA methods and generalizable to other learning setups such as semi-supervised learning, unsupervised model adaptation, etc.

show abstract

Uncertainty-Aware Unsupervised Domain Adaptation in Object Detection

Cited by 90 publications

References 75 publications

Unsupervised Domain Adaptation of Object Detectors: A Survey

Unsupervised Domain Adaptation of Object Detectors: A Survey

RDA: Robust Domain Adaptation via Fourier Adversarial Attacking

Category Contrast for Unsupervised Domain Adaptation in Visual Tasks

Contact Info

Product

Resources

About