Knowledge augmented visual learning
Loading...
Authors
Wang, Ziheng
Issue Date
2015-08
Type
Electronic thesis
Thesis
Thesis
Language
ENG
Keywords
Computer Systems engineering
Alternative Title
Abstract
To address these issues, we advocate that the long-term success of computer vision requires a union of data and knowledge, and knowledge is particularly important to complement data-driven approaches to achieve robust and generalizable performance. To this goal, we propose to develop methods to systemically identify and encode knowledge from different sources and to principally integrate it with the image data. This research includes three parts: knowledge taxonomy, knowledge encoding, and knowledge evaluation.
Description
August 2015
School of Engineering
School of Engineering
Full Citation
Publisher
Rensselaer Polytechnic Institute, Troy, NY