Joint Geometrical and Statistical Alignment for Visual Domain Adaptation

Abstract

This paper presents a novel unsupervised domain adaptation method for cross-domain visual recognition. We propose a unified framework that reduces the shift between domains both statistically and geometrically, referred to as Joint Geometrical and Statistical Alignment (JGSA). Specifically, we learn two coupled projections that project the source domain and target domain data into low dimensional subspaces where the geometrical shift and distribution shift are reduced simultaneously. The objective function can be solved efficiently in a closed form. Extensive experiments have verified that the proposed method significantly outperforms several state-of-the-art domain adaptation methods on a synthetic dataset and three different real world cross-domain visual recognition tasks.

References in corpus (2)

• Kernel Manifold Alignment
• Zero-Shot Domain Adaptation via Kernel Regression on the Grassmannian

Cited by in corpus (13)

• Deep Cocktail Network: Multi-source Unsupervised Domain Adaptation with Category Shift
• A Survey on Deep Transfer Learning
• RGB-D-based Human Motion Recognition with Deep Learning: A Survey
• Self-Supervised Training Enhances Online Continual Learning
• Cross-position Activity Recognition with Stratified Transfer Learning
• Importance Weighted Adversarial Nets for Partial Domain Adaptation
• Discriminative and Geometry Aware Unsupervised Domain Adaptation
• Discriminative Label Consistent Domain Adaptation
• Zero-Shot Deep Domain Adaptation
• Visual Domain Adaptation with Manifold Embedded Distribution Alignment
• Back to the Future -- Sequential Alignment of Text Representations
• Different Set Domain Adaptation for Brain-Computer Interfaces: A Label Alignment Approach
• Towards Visual Feature Translation