Kernelized Unified Domain Adaptation on Geometrical Manifolds
作者：
Highlights：
• Kernelized Unified Domain Adaptation on Geometrical Manifolds for Domain Adaptation is proposed.
• Use source domain labeled and target domain unlabeled data for training the model.
• Incorporate all the objectives that are essential for domain adaptation applications.
• Compared with various non-domain adaptation and the domain adaptation methods.
• Achieved much higher accuracies over open real-world PIE Face and Office Caltech datasets.
摘要
•Kernelized Unified Domain Adaptation on Geometrical Manifolds for Domain Adaptation is proposed.•Use source domain labeled and target domain unlabeled data for training the model.•Incorporate all the objectives that are essential for domain adaptation applications.•Compared with various non-domain adaptation and the domain adaptation methods.•Achieved much higher accuracies over open real-world PIE Face and Office Caltech datasets.
论文关键词：Manifold,Classification,Unsupervised learning,Domain Adaptation,Kernelization,Discriminant analysis,Transfer learning
论文评审过程：Received 25 June 2019, Revised 27 January 2020, Accepted 30 September 2020, Available online 16 October 2020, Version of Record 10 February 2021.
论文官网地址：https://doi.org/10.1016/j.eswa.2020.114078
