Domain Adaptation for Visual Applications: A Comprehensive Survey

By Gabriella Csurka. 2017

550 citations recorded by the source corpus.

The aim of this paper is to give an overview of domain adaptation and transfer learning with a specific view on visual applications. After a general motivation, we first position domain adaptation in the larger transfer learning problem. Second, we try to address and analyze briefly the state-of-the-art methods for different types of scenarios, first describing the historical shallow methods, addressing both the homogeneous and the heterogeneous domain adaptation methods. Third, we discuss the effect of the success of deep convolutional architectures which led to new type of domain adaptation methods that integrate the adaptation within the deep architecture. Fourth, we overview the methods that go beyond image categorization, such as object detection or image segmentation, video analyses or learning visual attributes. Finally, we conclude the paper with a section where we relate domain adaptation to other machine learning solutions.

Subfields: Artificial Intelligence, Computer Vision and Pattern Recognition, Radiology, Nuclear Medicine and Imaging

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

In Computer Vision and Pattern Recognition

In Radiology, Nuclear Medicine and Imaging

Citations in the corpus

Learning Robust Global Representations by Penalizing Local Predictive Power (2019, methodology): “For a general primer, we refer the reader to the following reviews (Weiss et al., 2016; Csurka, 2017; Wang and Deng, 2018).”
nocaps: novel object captioning at scale (2019, background): “Novel Object Captioning Novel object captioning includes aspects of both transfer learning and domain adaptation [6].”
Learning Robust Representations by Projecting Superficial Statistics Out (2018, methodology): “For a general primer, we refer the reader to these extensive reviews (Weiss et al., 2016; Csurka, 2017).”
Recognition in Terra Incognita (2018, methodology): “There have been many methods proposed for domain adaptation in classification [41].”
The limits and potentials of deep learning for robotics (2018, background): “To address this domain adaptation problem [23], [30], [83], a robotic vision system should be able to learn from new training samples of known classes during deployment and adopt its internal representations accordingly.”
Continual Learning for Robotics (2019, background): “In computer vision, transferring what has been learned from a past environment to new environments would be often referred to as domain adaptation (Patel et al., 2015; Csurka, 2017).”
Using Simulation and Domain Adaptation to Improve Efficiency of Deep Robotic Grasping (2018, background): “There has recently been a signiﬁcant amount of work on domain adaptation, particularly for computer vision [16], [17].”
Sim-To-Real via Sim-To-Sim: Data-Efficient Robotic Grasping via Randomized-To-Canonical Adaptation Networks (2018, background): “Visual domain adaptation [42, 13] is a process that allows a machine learning model trained with samples from a source domain to generalize to a target domain, by utilizing existing but (mostly) unlabeled target data.”
SpotTune: Transfer Learning Through Adaptive Fine-Tuning (2018, background): “There is a long history of transfer learning and domain adaptation methods in computer vision [8, 35].”
Implicit 3D Orientation Learning for 6D Object Detection from RGB Images (2018, background): “Domain Adaptation (DA) [5] refers to leveraging training data from a source domain to a target domain of which a small portion of labeled data (supervised DA) or unlabeled data (unsupervised DA) is available.”
Configurable 3D Scene Synthesis and 2D Image Rendering with Per-pixel Ground Truth Using Stochastic Grammars (2017, background): “A review of existing work in this area is beyond the scope of this paper; we refer the reader to a recent survey (Csurka 2017).”
A Survey of Unsupervised Deep Domain Adaptation (2018, background): “In some cases, prior surveys do not discuss domain mapping [48, 49, 121], normalization statistic-based [48, 49, 121, 285], or ensemble-based [48, 49, 121, 246, 285] methods.”
Sim2real transfer learning for 3D human pose estimation: motion to the rescue (2019, background): “These sim2real works in robotics build on a long tradition of domain adaptation for visual data, which can involve learning maps between feature spaces [21], invariant feature extractors [18], or image-to-image translation [98]; for a review, see [12, 54].”
Small Sample Learning in Big Data Era (2018, background): “Readers can refer to (Csurka, 2017; Venkateswara et al., 2017) for more technical details.”
Advancing Medical Imaging Informatics by Deep Learning-Based Domain Adaptation (2020, methodology): “DL-based DA is achieved using various representation learning strategies such as aligning the domain distributions, learning a mapping between domains, separating normalization statistics, and ensemble-based approaches [32-34].”
Self-Challenging Improves Cross-Domain Generalization (2020, background): “, the distribution of the testing samples) [5, 33].”
Adaptive Risk Minimization: A Meta-Learning Approach for Tackling Group Shift (2020, methodology): “Many of these frameworks, such as domain adaptation [13], assume access to unlabeled test data at training time, which are often not readily available and can be quite challenging to obtain.”
Label Efficient Learning of Transferable Representations acrosss Domains and Tasks (2017, background): “Domain adaptation seeks to learn from related source domains a well performing model on target data distribution [4].”
VisDA: The Visual Domain Adaptation Challenge (2017, background): “While this problem has been studied extensively in the domain adaptation literature [8], progress has been limited by the lack of large-scale challenge benchmarks.”