Domain Adaptive Faster R-CNN for Object Detection in the Wild

By Yuhua Chen, Wen Li, Christos Sakaridis, Dengxin Dai, Luc Van Gool. 2018

1,565 citations recorded by the source corpus.

Object detection typically assumes that training and test data are drawn from an identical distribution, which, however, does not always hold in practice. Such a distribution mismatch will lead to a significant performance drop. In this work, we aim to improve the cross-domain robustness of object detection. We tackle the domain shift on two levels: 1) the image-level shift, such as image style, illumination, etc., and 2) the instance-level shift, such as object appearance, size, etc. We build our approach based on the recent state-of-the-art Faster R-CNN model, and design two domain adaptation components, on image level and instance level, to reduce the domain discrepancy. The two domain adaptation components are based on $$-divergence theory, and are implemented by learning a domain classifier in adversarial training manner. The domain classifiers on different levels are further reinforced with a consistency regularization to learn a domain-invariant region proposal network (RPN) in the Faster R-CNN model. We evaluate our newly proposed approach using multiple datasets including Cityscapes, KITTI, SIM10K, etc. The results demonstrate the effectiveness of our proposed approach for robust object detection in various domain shift scenarios.

Subfields: Artificial Intelligence, Computer Vision and Pattern Recognition, Radiology, Nuclear Medicine and Imaging

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

In Computer Vision and Pattern Recognition

In Radiology, Nuclear Medicine and Imaging

Citations in the corpus

Revisiting Self-Supervised Visual Representation Learning (2019, methodology): “a model trained on the ImageNet (ILSVRC-2012) dataset [38] can only recognize 1000 semantic categories or a model that was trained to perceive road traffic at daylight may not work in darkness [5, 4].”
Benchmarking Robustness in Object Detection: Autonomous Driving when Winter is Coming (2019, background): “Evaluating robustness to environmental changes in autonomous driving In recent years, weather conditions turned out to be a central limitation for state-of-the art autonomous driving systems [Sakaridis et al., 2018a, Volk et al., 2019, Dai and Van Gool, 2018, Chen et al., 2018, Lee et al., 2018].”
ADVENT: Adversarial Entropy Minimization for Domain Adaptation in Semantic Segmentation (2018, methodology): “We conducted experiments in the UDA object detection set-up Cityscapes → Cityscapes-Foggy, similar to the one in [3].”
Self-Similarity Grouping: A Simple Unsupervised Cross Domain Adaptation Approach for Person Re-Identification (2018, background): “However, the traditional UDA approaches [4, 5, 27] always have an assumption that the source and target domain share the same set of classes, which does not hold for the person re-ID problem.”
Automatic Adaptation of Object Detectors to New Domains Using Self-Training (2019, methodology): “(a) Baseline; (b) HP [25]; (c) Ours; (d) DA[8].”
Adapting Object Detectors via Selective Cross-Domain Alignment (2019, methodology): “[4] 38.5 Ours-Type2( K =4; m =128;256 × 256) 41.9 Ours-Type3( K =8; m =64;128 × 128) 42.5 Table 4.”
Semi-Supervised Domain Adaptation via Minimax Entropy (2019, methodology): “UDA has been applied to various applications such as image classification [27], semantic segmentation [32], and object detection [6, 29].”
A Survey of Deep Learning-Based Object Detection (2019, methodology): “[277] utilize Faster R-CNN with a domain classifier trained to”
Exploring Categorical Regularization for Domain Adaptive Object Detection (2020, methodology): “Thanks to the flexibility of Faster R-CNN, recently, it is widely adapted for domain adaptive object detection [1, 28, 41, 12] with adversarial training [5].”
Exploring Object Relation in Mean Teacher for Cross-Domain Detection (2019, methodology): “Similar to previous work [4], our approach aims to leverage additional unlabeled target data for learning domain-invariant detector for cross-domain detection.”
Few-Shot Object Detection With Attention-RPN and Multi-Relation Detector (2019, methodology): “DA Faster R-CNN [66] uses massively annotated data from source domains (KITTI/Cityscapes) and unlabeled data from target domains (Cityscapes/KITTI) to train the domain adaptive Faster R-CNN, and evaluated the performance on target domains.”
Source Data-Absent Unsupervised Domain Adaptation Through Hypothesis Transfer and Labeling Transfer (2020, methodology): “nition [2], [4], [5], [6], semantic segmentation [3], [7], [8], [9], sentiment classification [10], [11], object detection [12], [13], and person re-identification [14], [15].”
Image-Adaptive YOLO for Object Detection in Adverse Weather Conditions (2021, methodology): “This problem can also be treated as an unsupervised domain adaptation task (Chen et al. 2018; Hnewa and Radha 2021).”
Multiple Class Novelty Detection Under Data Distribution Shift (2020, background): “However, most of the work on this topic has been done for the task of classiﬁcation [52,49,11,14], segmentation [50,55,15], detection [16,9,19] etc. and to the best of our knowledge no work is available in the literature that addresses the distribution shift problem for novelty detection.”
Cross-Domain Adaptive Teacher for Object Detection (2021, methodology): “Following [3] and [32], we employ Faster RCNN [30] as the base detection model in our Adaptive Teachera and implement it using Detectron2.”
Unsupervised Domain Adaptation of Object Detectors: A Survey (2021, methodology): “Multiple variations of this formulation that are commonly studied in the literature include, semi-supervised [96], [97], [98], weakly-supervised [99], and unsupervised domain adaptation [21], [26], [58], [73], [74], [100], [101].”
End-to-End Learning of Driving Models with Surround-View Cameras and Route Planners (2018, background): “Typical examples include the detection of roads [62], traffic lights [63], cars and pedestrians [64,65,1,2], and the tracking of such objects [66,67,68].”
Contrastive Test-Time Adaptation (2022, background): “Domain adaptation has been extensively explored for many visual tasks, including image classification [54], semantic segmentation [52] and object detection [9].”