Image to Image Translation for Domain Adaptation

Murez, Zak; Kolouri, Soheil; Kriegman, David; Ramamoorthi, Ravi; Kim, Kyungnam

doi:10.1109/cvpr.2018.00473

Cited by 470 publications

(302 citation statements)

References 31 publications

(89 reference statements)

Supporting

Mentioning

302

Contrasting

Order By: Relevance

“…Tab. 1 summarizes the differences in losses and network architecture between the proposed method and the baselines, as well as a recent domain adaptation work (a different but related task) called I2I work [20].…”

Section: A Discussion Of the Loss Termsmentioning

confidence: 99%

“…While the supervised case, where the training set consists of matching samples of input/output images, is of considerable practical interest [10,25], in many cases, such samples are very challenging to collect. Unsupervised domain translation methods receive a training set of unmatched samples, one set of samples from each domain, and learn to map between a sample in one domain and the analogous sample in the other domain [30,12,27,1,17,16,4,5,28,29,13,9,20].…”

Section: Related Workmentioning

confidence: 99%

“…where λ i are trade-off parameters and the additional loss terms are defined as: CycleGAN [30] MUNIT [9] disentangled disentangled OST [2] I2I [20] Ours unified Table 1. Comparison of losses and network architecture between the proposed method and our baselines.…”

Section: Phase II Of Trainingmentioning

confidence: 99%

“…Feature cycle A (resp. B) is the feature cycle loss for images encoded from domain A. GAN Z is the loss used by [20] for domain confusion in the latent space. GAN A (resp.…”

Section: Phase II Of Trainingmentioning

confidence: 99%

See 3 more Smart Citations

Bidirectional One-Shot Unsupervised Domain Mapping

Cohen

Wolf

2019

2019 IEEE/CVF International Conference on Computer Vision (ICCV)

View full text Add to dashboard Cite

We study the problem of mapping between a domain A, in which there is a single training sample and a domain B, for which we have a richer training set. The method we present is able to perform this mapping in both directions. For example, we can transfer all MNIST images to the visual domain captured by a single SVHN image and transform the SVHN image to the domain of the MNIST images. Our method is based on employing one encoder and one decoder for each domain, without utilizing weight sharing. The autoencoder of the single sample domain is trained to match both this sample and the latent space of domain B. Our results demonstrate convincing mapping between domains, where either the source or the target domain are defined by a single sample, far surpassing existing solutions. Our code is made publicly available at

show abstract

Section: A Discussion Of the Loss Termsmentioning

confidence: 99%

Section: Related Workmentioning

confidence: 99%

Section: Phase II Of Trainingmentioning

confidence: 99%

“…Feature cycle A (resp. B) is the feature cycle loss for images encoded from domain A. GAN Z is the loss used by [20] for domain confusion in the latent space. GAN A (resp.…”

Section: Phase II Of Trainingmentioning

confidence: 99%

See 2 more Smart Citations

Bidirectional One-Shot Unsupervised Domain Mapping

Cohen

Wolf

2019

2019 IEEE/CVF International Conference on Computer Vision (ICCV)

View full text Add to dashboard Cite

show abstract

“…They also extend their model with a multi-level adaptation framework to adapt features at different scales. Similarly, Murez et al [28] design an adaptation framework based on a backbone encoder supported by multiple auxiliary networks and losses with regularization purposes to achieve domain agnostic feature extraction. The discovered latent feature space is then exploited to learn a domain invariant predictor.…”

Section: Related Workmentioning

confidence: 99%

Unsupervised domain adaptation for mobile semantic segmentation based on cycle consistency and feature alignment

Toldo

Michieli

Agresti

et al. 2020

Image and Vision Computing

View full text Add to dashboard Cite

The supervised training of deep networks for semantic segmentation requires a huge amount of labeled real world data. To solve this issue, a commonly exploited workaround is to use synthetic data for training, but deep networks show a critical performance drop when analyzing data with slightly different statistical properties with respect to the training set. In this work, we propose a novel Unsupervised Domain Adaptation (UDA) strategy to address the domain shift issue between real world and synthetic representations. An adversarial model, based on the cycle consistency framework, performs the mapping between the synthetic and real domain. The data is then fed to a MobileNet-v2 architecture that performs the semantic segmentation task. An additional couple of discriminators, working at the feature level of the MobileNet-v2, allows to better align the features of the two domain distributions and to further improve the performance. Finally, the consistency of the semantic maps is exploited. After an initial supervised training on synthetic data, the whole UDA architecture is trained end-to-end considering all its components at once. Experimental results show how the proposed strategy is able to obtain impressive performance in adapting a segmentation network trained on synthetic data to real world scenarios. The usage of the lightweight MobileNet-v2 architecture allows its deployment on devices with limited computational resources as the ones employed in autonomous vehicles.

show abstract

Row‐sensing templates: A generic 3D sensor‐based approach to robot localization with respect to orchard row centerlines

Fei

Vougioukas

2022

Journal of Field Robotics

View full text Add to dashboard Cite

Accurate robot localization relative to orchard row centerlines is essential for autonomous guidance where satellite signals are often obstructed by foliage.Existing sensor-based approaches rely on various features extracted from images and point clouds. However, any selected features are not available consistently, because the visual and geometrical characteristics of orchard rows change drastically when tree types, growth stages, canopy management practices, seasons, and weather conditions change. In this study, we introduce a novel localization method that does not rely on features; instead, it relies on the concept of a row-sensing template, which is the expected observation of a 3D sensor traveling in an orchard row, when the sensor is anywhere on the centerline and perfectly aligned with it. First, the template is built using a few measurements, provided that the sensor's true pose with respect to the centerline is available. Then, during navigation, the best pose estimate (and its confidence) is estimated by maximizing the match between the template and the sensed point cloud using particle-filtering. The method can adapt to various orchards and conditions by rebuilding the template. Experiments were performed in a vineyard, and in an orchard in different seasons. Results showed that the lateral mean absolute error (MAE) was less than 3.6% of the row width, and the heading MAE was less than 1.72°. Localization was robust, as errors did not increase when less than 75% of measurement points were missing. The results indicate that template-based localization can provide a generic approach for accurate and robust localization in real-world orchards.

show abstract

Image to Image Translation for Domain Adaptation

Cited by 470 publications

References 31 publications

Bidirectional One-Shot Unsupervised Domain Mapping

Bidirectional One-Shot Unsupervised Domain Mapping

Unsupervised domain adaptation for mobile semantic segmentation based on cycle consistency and feature alignment

Row‐sensing templates: A generic 3D sensor‐based approach to robot localization with respect to orchard row centerlines

Contact Info

Product

Resources

About