Zero-Shot Domain Adaptation with a Physics Prior
- URL: http://arxiv.org/abs/2108.05137v1
- Date: Wed, 11 Aug 2021 10:28:56 GMT
- Title: Zero-Shot Domain Adaptation with a Physics Prior
- Authors: Attila Lengyel and Sourav Garg and Michael Milford and Jan C. van
Gemert
- Abstract summary: A traditional domain adaptation setting is to train on one domain and adapt to the target domain by exploiting unlabeled data samples from the test set.
We exploit a visual inductive prior derived from physics-based reflection models for domain adaptation.
We show that the color invariant layer reduces the day-night distribution shift in feature map activations throughout the network.
- Score: 39.424545456601074
- License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
- Abstract: We explore the zero-shot setting for day-night domain adaptation. The
traditional domain adaptation setting is to train on one domain and adapt to
the target domain by exploiting unlabeled data samples from the test set. As
gathering relevant test data is expensive and sometimes even impossible, we
remove any reliance on test data imagery and instead exploit a visual inductive
prior derived from physics-based reflection models for domain adaptation. We
cast a number of color invariant edge detectors as trainable layers in a
convolutional neural network and evaluate their robustness to illumination
changes. We show that the color invariant layer reduces the day-night
distribution shift in feature map activations throughout the network. We
demonstrate improved performance for zero-shot day to night domain adaptation
on both synthetic as well as natural datasets in various tasks, including
classification, segmentation and place recognition.
Related papers
- Learning Visual Conditioning Tokens to Correct Domain Shift for Fully Test-time Adaptation [24.294049653744185]
In transformer-based image classification, the class token at the first transformer encoder layer can be learned to capture the domain-specific characteristics of target samples during test-time adaptation.
We propose a bi-level learning approach to capture the long-term variations of domain-specific characteristics while accommodating local variations of instance-specific characteristics.
Our proposed bi-level visual conditioning token learning method is able to achieve significantly improved test-time adaptation performance by up to 1.9%.
arXiv Detail & Related papers (2024-06-27T17:16:23Z) - ShadowSense: Unsupervised Domain Adaptation and Feature Fusion for
Shadow-Agnostic Tree Crown Detection from RGB-Thermal Drone Imagery [7.2038295985918825]
This paper presents a novel method for detecting shadowed tree crowns from remote sensing data.
The proposed method (ShadowSense) is entirely self-supervised, leveraging domain adversarial training without source domain annotations.
It then fuses complementary information of both modalities to effectively improve upon the predictions of an RGB-trained detector.
arXiv Detail & Related papers (2023-10-24T22:01:14Z) - Compositional Semantic Mix for Domain Adaptation in Point Cloud
Segmentation [65.78246406460305]
compositional semantic mixing represents the first unsupervised domain adaptation technique for point cloud segmentation.
We present a two-branch symmetric network architecture capable of concurrently processing point clouds from a source domain (e.g. synthetic) and point clouds from a target domain (e.g. real-world)
arXiv Detail & Related papers (2023-08-28T14:43:36Z) - Similarity Min-Max: Zero-Shot Day-Night Domain Adaptation [52.923298434948606]
Low-light conditions not only hamper human visual experience but also degrade the model's performance on downstream vision tasks.
This paper challenges a more complicated scenario with border applicability, i.e., zero-shot day-night domain adaptation.
We propose a similarity min-max paradigm that considers them under a unified framework.
arXiv Detail & Related papers (2023-07-17T18:50:15Z) - Domain Adaptive Object Detection for Autonomous Driving under Foggy
Weather [25.964194141706923]
This paper proposes a novel domain adaptive object detection framework for autonomous driving under foggy weather.
Our method leverages both image-level and object-level adaptation to diminish the domain discrepancy in image style and object appearance.
Experimental results on public benchmarks show the effectiveness and accuracy of the proposed method.
arXiv Detail & Related papers (2022-10-27T05:09:10Z) - Feather-Light Fourier Domain Adaptation in Magnetic Resonance Imaging [2.024988885579277]
Generalizability of deep learning models may be severely affected by the difference in the distributions of the train (source domain) and the test (target domain) sets.
We propose a very light and transparent approach to perform test-time domain adaptation.
arXiv Detail & Related papers (2022-07-31T17:28:42Z) - A Comprehensive Study of Image Classification Model Sensitivity to
Foregrounds, Backgrounds, and Visual Attributes [58.633364000258645]
We call this dataset RIVAL10 consisting of roughly $26k$ instances over $10$ classes.
We evaluate the sensitivity of a broad set of models to noise corruptions in foregrounds, backgrounds and attributes.
In our analysis, we consider diverse state-of-the-art architectures (ResNets, Transformers) and training procedures (CLIP, SimCLR, DeiT, Adversarial Training)
arXiv Detail & Related papers (2022-01-26T06:31:28Z) - Semi-Supervised Domain Adaptation with Prototypical Alignment and
Consistency Learning [86.6929930921905]
This paper studies how much it can help address domain shifts if we further have a few target samples labeled.
To explore the full potential of landmarks, we incorporate a prototypical alignment (PA) module which calculates a target prototype for each class from the landmarks.
Specifically, we severely perturb the labeled images, making PA non-trivial to achieve and thus promoting model generalizability.
arXiv Detail & Related papers (2021-04-19T08:46:08Z) - Synthetic-to-Real Domain Adaptation for Lane Detection [5.811502603310248]
We explore learning from abundant, randomly generated synthetic data, together with unlabeled or partially labeled target domain data.
This poses the challenge of adapting models learned on the unrealistic synthetic domain to real images.
We develop a novel autoencoder-based approach that uses synthetic labels unaligned with particular images for adapting to target domain data.
arXiv Detail & Related papers (2020-07-08T10:54:21Z) - Phase Consistent Ecological Domain Adaptation [76.75730500201536]
We focus on the task of semantic segmentation, where annotated synthetic data are aplenty, but annotating real data is laborious.
The first criterion, inspired by visual psychophysics, is that the map between the two image domains be phase-preserving.
The second criterion aims to leverage ecological statistics, or regularities in the scene which are manifest in any image of it, regardless of the characteristics of the illuminant or the imaging sensor.
arXiv Detail & Related papers (2020-04-10T06:58:03Z)
This list is automatically generated from the titles and abstracts of the papers in this site.
This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.