Cross-domain learning methods for high-level visual concept classification

By Wei Jiang, Eric Zavesky, Shih-Fu Chang, Alexander C. Loui. 2008

186 citations recorded by the source corpus.

Subfields: Artificial Intelligence, Computer Vision and Pattern Recognition

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

Citations in the corpus

Efficient Learning of Domain-invariant Image Representations (2013, neutral): “In the linear case, feature replication [17] can be shown to decompose the learned parameter into θ = θ̂ + θ′, where θ̂ is shared by all domains [22], in a similar fashion to adaptive SVMs.”
Domain Adaptation for Visual Applications: A Comprehensive Survey (2017, background): “The cross-domain SVM, proposed in [58], constrains the impact of source data to the k-nearest neighbors (similarly to the spirit of the Localized SVM [59]).”
Data-intensive applications, challenges, techniques and technologies: A survey on Big Data (2014, background): “There are many scale machine learning algorithms [152,177,106,85], but many important specific sub-fields in large-scale machine learning, such as large-scale recommender systems, natural language processing, association rule learning, ensemble learning, still face the scalability problems.”
Human Reidentification with Transferred Metric Learning (2012, methodology): “However, different than [24], a distance metric instead of a hyperplane is learned in our case.”
Concept-Based Video Retrieval (2009, background): “concept detector to the domain of interest by adjusting the classifier [108, 289], or by exploiting massive sets of image examples obtained from the (social-tagged) web [131, 238, 260] in combination with semisupervised learning [37, 303].”
Cross-Domain Object Recognition Via Input-Output Kernel Analysis (2013, methodology): “2) A-SVM applies Adaptive SVM [5] on the training data from both domains.”
Generalized Transfer Subspace Learning Through Low-Rank Constraint (2014, background): “More and more research effort is being devoted to visual domain adaptation (Yang et al. 2007; Jiang et al. 2008; Wang et al. 2008; Duan et al. 2009, 2012a,b,c; Si et al. 2010; Saenko et al. 2010; Gopalan et al. 2011; Kulis et al. 2011; Aytar and Zisserman 2011; Qi et al. 2011; Lim et al. 2011; Gong ”
Where am I in the dark: Exploring active transfer learning on the use of indoor localization based on thermal imaging (2016, methodology): “We compare our active transfer learning method with other methods (SVM trained with thermal images, SVM trained with visible images, Active SVM trained by thermal images, Active SVM trained by visible images, cross-domain SVM (CD-SVM) [18], active transfer domain knowledge [19] (ATDK), multiple kern”
Adapting Visual Category Models to New Domains (2010, background): “[13] defines a weight for each source training sample based on distance to the target domain, and re-trains the SVM classifier with re-weighted patterns.”
Domain Transfer SVM for video concept detection (2009, methodology): “Following this work, cross-domain SVM (CD-SVM) proposed by Jiang et al. [7] used k-nearest neighbors from the target domain to define a weight for each auxiliary pattern, and then the SVM classifier was trained with re-weighted patterns.”
Boosting for transfer learning with multiple sources (2010, background): “[14] derived more adaptable decision boundaries by training a target SVM with the help of weighted support vectors learned from multiple sources.”
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”
Domain Transfer Multiple Kernel Learning (2012, methodology): “…cross-domain learning methods have been successfully used in many real-world applications, such as sentiment classification [2], natural language processing [11], text categorization [9], [21], information extraction [9], WiFi localization [21], and visual concept classification [16], [17], [36].”