Domain Adaptation for Structured Output via Discriminative Patch Representations

@article{Tsai2019DomainAF,
  title={Domain Adaptation for Structured Output via Discriminative Patch Representations},
  author={Yi-Hsuan Tsai and Kihyuk Sohn and Samuel Schulter and Manmohan Chandraker},
  journal={2019 IEEE/CVF International Conference on Computer Vision (ICCV)},
  year={2019},
  pages={1456-1465}
}
Predicting structured outputs such as semantic segmentation relies on expensive per-pixel annotations to learn supervised models like convolutional neural networks. However, models trained on one data domain may not generalize well to other domains without annotations for model finetuning. To avoid the labor-intensive process of annotation, we develop a domain adaptation method to adapt the source data to the unlabeled target domain. We propose to learn discriminative feature representations of… 
Contrastive Learning and Self-Training for Unsupervised Domain Adaptation in Semantic Segmentation
TLDR
This work proposes a contrastive learning approach that adapts category-wise centroids across domains and extends it with selftraining, where it uses a memory-efficient temporal ensemble to generate consistent and reliable pseudo-labels.
Geometric Unsupervised Domain Adaptation for Semantic Segmentation
TLDR
Geometric Unsupervised Domain Adaptation method (GUDA) establishes a new state of the art in UDA for semantic segmentation on three benchmarks, outperforming methods that use domain adversarial learning, self-training, or other self-supervised proxy tasks.
Variational Autoencoder Based Unsupervised Domain Adaptation For Semantic Segmentation
TLDR
A novel method that employs a variational autoencoder to get to the same destination but in a non-adversarial manner for unsupervised domain adaptation and integrates an adversarial loss into this method.
Unsupervised Domain Adaptation for Semantic Segmentation with Symmetric Adaptation Consistency
TLDR
A new approach is proposed that trains two segmentation models with the adversarial learning symmetrically and further introduces the consistency between the outputs of the two models into the semi-supervised learning to improve the accuracy of pseudo labels which significantly affect the final adaptation performance.
Adapting Segmentation Networks to New Domains by Disentangling Latent Representations
TLDR
This work designs and carefully analyze multiple latent space-shaping regularization strategies that work in conjunction to reduce the domain discrepancy in semantic segmentation and proposes a novel performance metric to capture the relative efficacy of an adaptation strategy compared to supervised training.
Joint Adversarial Learning for Domain Adaptation in Semantic Segmentation
TLDR
This work proposes a novel joint adversarial learning (JAL) to boost the domain discriminator in output space by introducing the information ofdomain discriminator from low-level features and proposes a weight transfer module (WTM) to alleviate the inherent bias of the trained decoder towards source domain.
PixMatch: Unsupervised Domain Adaptation via Pixelwise Consistency Training
TLDR
This work presents a novel framework for unsupervised domain adaptation based on the notion of target-domain consistency training, and introduces a new loss term to enforce pixelwise consistency between the model’s predictions on a target image and a perturbed version of the same image.
Unsupervised Domain Adaptation with Multiple Domain Discriminators and Adaptive Self-Training
TLDR
Experimental results prove the effectiveness of the proposed strategy in adapting a segmentation network trained on synthetic datasets like GTA5 and SYNTHIA, to real world datasets like Cityscapes and Mapillary.
Unsupervised Domain Adaptation in Semantic Segmentation via Orthogonal and Clustered Embeddings
TLDR
An effective Unsupervised Domain Adaptation (UDA) strategy, based on a feature clustering method that captures the different semantic modes of the feature distribution and groups features of the same class into tight and well-separated clusters is proposed.
Generating Target Image-Label Pairs for Unsupervised Domain Adaptation
TLDR
This work proposes an unsupervised domain adaptation method by generating image-label pairs in the target domain, in which the model is augmented with the generated target pairs and achieve class-level transfer.
...
...

References

SHOWING 1-10 OF 62 REFERENCES
Learning to Adapt Structured Output Space for Semantic Segmentation
TLDR
A multi-level adversarial network is constructed to effectively perform output space domain adaptation at different feature levels and it is shown that the proposed method performs favorably against the state-of-the-art methods in terms of accuracy and visual quality.
Domain Transfer Through Deep Activation Matching
TLDR
A layer-wise unsupervised domain adaptation approach for semantic segmentation that uses a Generative Adversarial Network (or GAN) to align activation distributions and achieves state-of-the-art results on a variety of popular domain adaptation tasks.
Unsupervised Pixel-Level Domain Adaptation with Generative Adversarial Networks
TLDR
This generative adversarial network (GAN)-based method adapts source-domain images to appear as if drawn from the target domain, and outperforms the state-of-the-art on a number of unsupervised domain adaptation scenarios by large margins.
FCNs in the Wild: Pixel-level Adversarial and Constraint-based Adaptation
TLDR
This paper introduces the first domain adaptive semantic segmentation method, proposing an unsupervised adversarial approach to pixel prediction problems, and outperforms baselines across different settings on multiple large-scale datasets.
Unsupervised Domain Adaptation for Semantic Segmentation via Class-Balanced Self-training
TLDR
This paper proposes a novel UDA framework based on an iterative self-training (ST) procedure, where the problem is formulated as latent variable loss minimization, and can be solved by alternatively generating pseudo labels on target data and re-training the model with these labels.
Gotta Adapt 'Em All: Joint Pixel and Feature-Level Domain Adaptation for Recognition in the Wild
TLDR
This work uses 3D geometry and image synthesis based on a generalized appearance flow to preserve identity across pose transformations, while using an attribute-conditioned CycleGAN to translate a single source into multiple target images that differ in lower-level properties such as lighting.
Image to Image Translation for Domain Adaptation
TLDR
This work proposes the novel use of the recently proposed unpaired image-to-image translation framework to constrain the features extracted by the backbone encoder network, and applies it to domain adaptation between MNIST, USPS, and SVHN datasets, and Amazon, Webcam and DSLR Office datasets in classification tasks, and also between GTA5 and Cityscapes datasets for a segmentation task.
Unsupervised Domain Adaptation with Residual Transfer Networks
TLDR
Empirical evidence shows that the new approach to domain adaptation in deep networks that can jointly learn adaptive classifiers and transferable features from labeled data in the source domain and unlabeledData in the target domain outperforms state of the art methods on standard domain adaptation benchmarks.
Penalizing Top Performers: Conservative Loss for Semantic Segmentation Adaptation
TLDR
This work proposes a novel loss function, i.e., Conservative Loss, which penalizes the extreme good and bad cases while encouraging the moderate examples and enables the network to learn features that are discriminative by gradient descent and are invariant to the change of domains via gradient ascend method.
No More Discrimination: Cross City Adaptation of Road Scene Segmenters
TLDR
This work proposes an unsupervised learning approach to adapt road scene segmenters across different cities by advancing a joint global and class-specific domain adversarial learning framework, and shows that this method improves the performance of semantic segmentation in multiple cities across continents.
...
...