What question did this study set out to answer?

This study aims to recognize RGB images and videos by leveraging additional depth information from RGB-D data in a new UDA framework.

Visual Recognition in RGB Images and Videos by Learning from RGB-D Data

View Full Paper

Ask AI

Bookmark

Authors

Randomized trial demonstrates improved image and video recognition using RGB-D depth data, suggesting effective learning strategies.

This study aims to recognize RGB images and videos by leveraging additional depth information from RGB-D data in a new UDA framework.
Formulate unsupervised domain adaptation (UDA) to address data distribution mismatch between source and target domains.
Develop a unified learning framework called DAM2S, incorporating various regularizers and strategies for robust classifier learning.
Conduct experiments on object recognition and cross-view action recognition to test the effectiveness of the proposed methods.
Achieved significant improvement in recognition accuracy for RGB images and videos leveraging RGB-D data.
Effectiveness of strategies incorporated in DAM2S framework demonstrated through various controlled experiments.
Enhanced classification performance in cross-dataset and cross-view scenarios indicating the utility of depth features.

Li et al. (2017) studied this question.

View Full Paper

Ask AI

Bookmark