We present the 2017 Visual Domain Adaptation (VisDA) dataset and challenge1, a large-scale testbed for unsupervised domain adaptation across visual domains. Unsupervised domain adaptation aims to solve the real-world problem of domain shift, where machine learning models trained on one domain must be transferred and adapted to a novel visual domain without additional supervision. The VisDA2017 ...

Deep neural networks are able to learn powerful representations from large quantities of labeled input data, however they cannot always generalize well across changes in input distributions. Domain adaptation algorithms have been proposed to compensate for the degradation in performance due to domain shift. In this paper, we address the case when the target domain is unlabeled, requiring unsupe...

Reliable perception during fast motion maneuvers or in high dynamic range environments is crucial for robotic systems. Since event cameras are robust to these challenging conditions, they have great potential increase the reliability of robot vision. However, event-based vision has been held back by shortage labeled datasets due novelty cameras. To overcome this drawback, we propose a task tran...

Unsupervised domain adaptation (UDA) transfers knowledge from a label-rich source to different but related fully-unlabeled target domain. To address the problem of shift, more and UDA methods adopt pseudo labels samples improve generalization ability on However, inaccurate may yield suboptimal performance with error accumulation during optimization process. Moreover, once are generated, how rem...

Source-Free Unsupervised Domain Adaptation(SFUDA) aims to adapt a pre-trained source model an unlabeled target domain without access the original labeled samples. Many existing SFUDA approaches apply self-training strategy, which involves iteratively selecting confidently predicted samples as pseudo-labeled used train fit domain. However, strategy may also suffer from sample selection bias and ...

In this paper, we quest the capability of transferring quality natural scene images to that are not acquired by optical cameras (e.g., screen content images, SCIs), rooted in widely accepted view human visual system has adapted and evolved through perception environment. Here, develop first unsupervised domain adaptation based no reference assessment method for SCIs, leveraging rich subjective ...

This paper contributes to improving a bottleneck residual block-based feature extractor as set of layers for transforming raw data into features classification. structure is utilized avoid the issues deep learning network, such overfitting problems and low computational efficiency caused by redundant computation, high dimensionality, gradient vanishing. With this structure, domain adversarial n...

Recent advances in unsupervised domain adaptation mainly focus on learning shared representations by global statistics alignment, such as the Maximum Mean Discrepancy (MMD) which matches across domains. The lack of class information, however, may lead to partial alignment (or even misalignment) and poor generalization performance. For robust we argue that similarities different features source ...

In recent years, supervised person re-identification (re-ID) models have received increasing studies. However, these trained on the source domain always suffer dramatic performance drop when tested an unseen domain. Existing methods are primary to use pseudo labels alleviate this problem. One of most successful approaches predicts neighbors each unlabeled image and then uses them train model. A...