dc.contributor.advisor | Potetz, Brian | |
dc.contributor.author | Kroge, Jason | |
dc.date.accessioned | 2010-06-09T03:52:24Z | |
dc.date.available | 2010-06-09T03:52:24Z | |
dc.date.issued | 2010-04-28 | |
dc.date.submitted | 2010 | |
dc.identifier.other | http://dissertations.umi.com/ku:10898 | |
dc.identifier.uri | http://hdl.handle.net/1808/6294 | |
dc.description.abstract | With the shear amount and variety of digital images available in the world today, people need an effective method to search for any particular image. The commonly used strategy of searching by keyword has several problems, especially when searching for aspects that are difficult to describe with words. In this paper, I will discuss an image retrieval system that can be used to search for visually-similar images based on image content rather than associated keywords. I will discuss the major components of this system including a pre-processing step using Haar wavelets and the steps for training a deep belief network to recognize higher-order features that may have a semantic or category specific meaning. The paper concludes with a comparison of performance between the newly proposed system and other published results. | |
dc.format.extent | 26 pages | |
dc.language.iso | EN | |
dc.publisher | University of Kansas | |
dc.rights | This item is protected by copyright and unless otherwise specified the copyright of this thesis/dissertation is held by the author. | |
dc.subject | Computer science | |
dc.title | Content-Based Image Retrieval Using Deep Belief Networks | |
dc.type | Thesis | |
dc.contributor.cmtemember | Chen, Xue-wen | |
dc.contributor.cmtemember | Luo, Bo | |
dc.thesis.degreeDiscipline | Electrical Engineering & Computer Science | |
dc.thesis.degreeLevel | M.S. | |
kusw.oastatus | na | |
kusw.oapolicy | This item does not meet KU Open Access policy criteria. | |
kusw.bibid | 7078809 | |
dc.rights.accessrights | openAccess | |