Learning Image Manifolds by Semantic Subspace Projection

dc.contributor.authorYu, Jie
dc.contributor.authorTian, Qi
dc.date.accessioned2023-10-23T15:09:39Z
dc.date.available2023-10-23T15:09:39Z
dc.date.issued2006-10
dc.description.abstractIn many image retrieval applications, the mapping between high-level semantic concept and low-level features is obtained through a learning process. Traditional approaches often assume that images with same semantic label share strong visual similarities and should be clustered together to facilitate modeling and classification. Our research indicates this assumption is inappropriate in many cases. Instead we model the images as lying on non-linear image subspaces embedded in the high-dimensional space and find that multiple subspaces may correspond to one semantic concept. By intelligently utilizing the similarity and dissimilarity information in semantic and geometric (image) domains, we find an optimal Semantic Subspace Projection (SSP) that captures the most important properties of the subspaces with respect to classification. Theoretical analysis proves that the well-known Linear Discriminant Analysis (LDA) could be formulated as a special case of our proposed method. To capture the semantic concept dynamically, SSP can integrate relevance feedback efficiently through incremental learning. Extensive experiments have been designed and conducted to compare our proposed method to the state-of-the-art techniques such as LDA, Locality Preservation Projection (LPP), Local Linear Embedding (LLE), Local Discriminant Embedding (LDE) and their semi-supervised algorithms. The results show the superior performance of SSP.
dc.description.departmentComputer Science
dc.description.sponsorshipThis work was supported in part by the Army Research Office (ARO) grant under W911NF-05-1-0404, and by the Center of Infrastructure Assurance and Security (CIAS), the University of Texas at San Antonio.
dc.identifier.urihttps://hdl.handle.net/20.500.12588/2122
dc.language.isoen_US
dc.publisherUTSA Department of Computer Science
dc.relation.ispartofseriesTechnical Report; CS-TR-2006-011
dc.subjectalgorithms
dc.subjecttheory
dc.subjectperformance
dc.subjectexperimentation
dc.subjectmeasurement
dc.subjectsemantic subspace projection
dc.subjectimage retrieval
dc.subjectrelevance feedback
dc.subjectsubspace learning
dc.subjectprincipal component analysis
dc.subjectlinear discriminant analysis
dc.titleLearning Image Manifolds by Semantic Subspace Projection
dc.typeTechnical Report

Files

Original bundle

Now showing 1 - 1 of 1
Loading...
Thumbnail Image
Name:
Yu_Tian_CS-TR-2006-011.pdf
Size:
830.7 KB
Format:
Adobe Portable Document Format

License bundle

Now showing 1 - 1 of 1
No Thumbnail Available
Name:
license.txt
Size:
1.86 KB
Format:
Item-specific license agreed upon to submission
Description: