Domain adaptation from RGB-D to RGB images
详细信息 查看全文
- 作者：Xiao Li ; Min Fang ; mfang@mail.xidian.edu.cn" class="auth_mail" title="E-mail the corresponding author ; Ju-Jie Zhang ; Jinqiao Wu
- 关键词：Domain adaptation ; Object recognition ; Depth images ; Trace norm
- 刊名：Signal Processing
- 出版年：2017
- 出版时间：February 2017
- 年：2017
- 卷：131
- 期：Complete
- 页码：27-35
- 全文大小：1540 K
文摘
- listitem" id="list_li0005">
- •
We aim to use RGB-D images to help recognizing RGB images.
- •
The correlation between RGB and depth images in source domain is maximized.
- •
The cross domain and cross modal constraints are jointly incorporated in the model.
- •
A unified framework is presented to learn the classifier parameters.
- •
Extensive experiment results show that depth information is useful for DA problems.