Self-Adversarial Disentangling for Specific Domain Adaptation

  title={Self-Adversarial Disentangling for Specific Domain Adaptation},
  author={Qianyu Zhou and Qiqi Gu and Jiangmiao Pang and Zhengyang Feng and Guangliang Cheng and Xuequan Lu and Jianping Shi and Lizhuang Ma},
Domain adaptation aims to bridge the domain shifts between the source and target domains. These shifts may span different dimensions such as fog, rainfall, etc. However, recent methods typically do not consider explicit prior knowledge on a specific dimension, thus leading to less desired adaptation performance. In this paper, we study a practical setting called Specific Domain Adaptation (SDA) that aligns the source and target domains in a demanded-specific dimension. Within this setting, we… 

Generative Domain Adaptation for Face Anti-Spoofing

This work proposes a novel perspective of UDA FAS that directly fits the target data to the models, and feeds the stylized data into the well-trained source model for classification, and combines two carefully designed consistency constraints.

Domain Adaptation: Challenges, Methods, Datasets, and Applications

This survey paper discusses how DA can help DNNs work efficiently in these settings by reviewing DA methods and techniques, and presents a comprehensive taxonomy, including the methods, datasets, challenges, and applications corresponding to each domain.

Robust Domain Adaptive Object Detection with Unified Multi-Granularity Alignment

An adaptive exponential moving average (AEMA) strategy that explores model assessments for model update to improve pseudo labels and alleviate local misalignment problem, boosting detection robustness and validate the superiority of MGA over other approaches on FCOS and Faster R-CNN detectors.

Context-Aware Mixup for Domain Adaptive Semantic Segmentation

A novel Context-Aware Mixup (CAMix) framework for domain adaptive semantic segmentation, which exploits this important clue of context-dependency as explicit prior knowledge in a fully end-to-end trainable manner for enhancing the adapt- ability toward the target domain.

Prototypical Contrast Adaptation for Domain Adaptive Semantic Segmentation

The proposed Prototypical Contrast Adaptation (ProCA), a simple and efficient contrastive learning method for unsupervised domain adaptive semantic segmentation, achieves the state-of-the-art performance on challenging benchmarks, outperforming previous methods by a large margin.

Interpolating Compressed Parameter Subspaces

The utility of CPS is demonstrated for single and multiple test-time distribution settings, with improved mappings between the two spaces with higher accuracy, improved robustness performance across perturbation types, reduced catastrophic forgetting on Split-CIFAR10/100, strong capacity for multi-task solutions and unseen/distant tasks, and storage-efficient inference (ensembling, hypernetworks).

Domain Adaptive Semantic Segmentation via Regional Contrastive Consistency Regularization

A novel and fully end-to-end trainable approach, called regional contrastive consistency regularization (RCCR) for domain adaptive semantic segmentation, designed to learn more robust and stable region-wise features under varying environ-ments.

PIT: Position-Invariant Transform for Cross-FoV Domain Adaptation

The Position-Invariant Transform (PIT) is proposed to better align images in different domains and can soundly boost the performance on both cross-domain object detection and segmentation for state-of-the-art techniques.

Object Detection in Foggy Scenes by Embedding Depth and Reconstruction into Domain Adaptation

This work proposes an encoder-decoder framework to reconstruct the fog-free background image and involves the target data in training both the authors' DA module and their detection module in a semi-supervised manner, so that the detection module is also exposed to the unlabeled target data.

FEW-Shot Cross-Sensor Domain Adaptation Between SAR and Multispectral Data

A novel few-shot cross-sensor domain adaptation technique between SAR and multispectral data for LULC classification and proposes a novel Feature Domain Alignment (FDA) loss function to align higher dimension features between mult ispectral and SAR domain.