Domain Transfer SVM for video concept detection

By Lixin Duan, Ivor Wai-Hung Tsang, Dong Xu, S. Maybank. 2009

305 citations recorded by the source corpus.

Subfields: Artificial Intelligence, Computer Vision and Pattern Recognition, Representation & Self-Supervised Learning

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 Representation & Self-Supervised Learning

Citations in the corpus

Few-Shot Adversarial Domain Adaptation (2017, methodology): “Finally, some use regularization to improve the fit on the target domain [4, 1, 62, 10, 3, 8].”
Transductive Multi-View Zero-Shot Learning (2015, background): “This process alleviates the projection domain shift problem, as well as providing a common space in which heterogeneous views can be directly compared, and their complementarity exploited (Sec.”
Efficient Learning of Domain-invariant Image Representations (2013, methodology): “We implement the transform learning and then apply both a KNN classifier (as originally proposed) and an SVM classifier.”
Unified Deep Supervised Domain Adaptation and Generalization (2017, background): “The third one regularizes a classifier trained on a source distribution to work well on a target distribution [4, 1, 66, 15, 3, 12].”
Domain Adaptation for Visual Applications: A Comprehensive Survey (2017, methodology): “eature representation for the given the task. When this is possible, shallow DA methods described in the Section 3 can be applied to the problem. Hence, DA solutions such as Adaptive SVM [258], DT-SVM[68], Adaptive MKL [69] or Selective Transfer Machine [44], were applied to video concept detection ”
SpotTune: Transfer Learning Through Adaptive Fine-Tuning (2018, methodology): “er learning and domain adaptation methods in computer vision [9,37]. Early approaches have concentrated on shallow classiﬁers, using techniques such as instance re-weighting [50,12], model adaptation [11,52], and feature space alignment [44,35,18]. In the multi-task setting, knowing which tasks or p”
A Two-Stage Weighting Framework for Multi-Source Domain Adaptation (2011, methodology): “It has been applied successfully in various applications including text classification (parts of speech tagging, webpage tagging, etc) [2], video concept detection across different TV channels [3], sentiment analysis (identifying positive and negative reviews across domains) [4] and WiFi Localizatio”
Transfer defect learning (2013, background): “In terms of applications, transfer learning has been widely studied to address cross-domain problems in text classification [15], [53], [54], natural language processing [25], [50], WiFi-based indoor localization [55], and computer vision [56].”
Distribution Calibration in Riemannian Symmetric Space (2011, methodology): “For example, Pan et al. (2008) proposed a transductive dimension reduction algorithm, i.e., maximum mean discrepancy embedding (MMDE), to minimize MMD in a latent subspace for cross-domain text categorization.”
Cross-Domain Object Recognition Via Input-Output Kernel Analysis (2013, methodology): “Other approaches used in vision researches such as [14], [15 ], [21] employ a domain similarity measure based on MMD [18], which is non-parametric, easy to estimate empirically and a lso flexible for the choice of kernel functions.”
Multiple source domain adaptation: A sharper bound using weighted Rademacher complexity (2015, unknown): “For example: sentiment analysis [1- 978-1-4673-9606-6115/$31.00 ©2015 IEEE 546 4], ranking model [5], statistical machine translation [6], information extraction [7] [8], named entity recognition [9], computer vision [10].”
Adapting Visual Category Models to New Domains (2010, methodology): “[9] used multiple-kernel learning to minimize the difference between the means of the source and target feature distributions.”
Tabula rasa: Model transfer for object category detection (2011, background): “There is another school of transfer learning where classifiers are transferred between domains [3, 4, 20, 28], for example by learning feature distributions for the source and target domains, but we are not concerned with this type of domain transfer problem here.”
Sim-Real Joint Reinforcement Transfer for 3D Indoor Navigation (2019, methodology): “This kind of method, also called Domain Adaptation, is widely applied in numerous tasks [22, 32, 43, 23, 9, 38].”
Boosting for transfer learning with multiple sources (2010, background): “[6] performed video concept detection by learning an SVM where the kernel is derived by minimizing the distribution mismatch between the labeled source data and the unlabeled target data.”
Discovering Latent Domains for Multisource Domain Adaptation (2012, methodology): “For example, if the domain specific classifier is (kernelized) nearest neighbors, as in our experiments , and we let x * k (y) be the most similar point in domain k to test point y, then our domain specific object category probability can be expressed as: p(c|d = k, y) = φ A (x * k (y)) T W j φ B (y”
Small Sample Learning in Big Data Era (2018, methodology): “Along this research direction, cross-domain SVM (Jiang et al., 2008), domain transfer SVM (Duan et al., 2009), domain adaptation SVM (Bruzzone and Marconcini, 2010), adaptive multiple kernel learning (A-MKL) (Duan et al., 2012), and residual transfer network (RTN) (Long et al., 2016) have been progr”