[April 4th] Seminar talk by Ms. Yijuan Lu (University of Texas at San Antonio)
Discriminant Analysis and Concepts Learning for Large-scale Image Data Set
- Speaker: Ms. Yijuan Lu
- Location: CH 430
- Date: Friday, April 4
- Time: 11:00 am - 12:00 pm
Seminar Abstract
Image Retrieval has been continuously at the fore-front of innovation
in computer science and plays a very important role in Multimedia
Information Retrieval (MIR). The content extraction, indexing, and
retrieval of large scale image database continue to be one of the most
challenging and fast-growing research areas. Digital libraries,
education and training, media commerce, home media, bio-computing in
bioinformatics, biometrics, and medical multimedia database, have
created a worldwide need for new paradigms and techniques on how to
browse, search, and summarize image collections. However, image
retrieval still faces some challenges such as semantic gap, small
sample size and high dimensionality. In this talk, several novel
semantic computing and discriminant analysis methods are presented
from low-high manner to alleviate these problems.
In the low-level analysis, Adaptive Discriminant Analysis (ADA) is
developed for feature dimension reduction for classification. It
combines the strength of linear discriminant analysis (LDA) and biased
discriminant analysis (BDA) by controlling the scatterness and
clusterness among samples from different classes, which allows it to
handle imbalanced data sets adaptively. In the middle-level analysis,
interactive boosting (i.Boosting) is further explored to enhance the
ADA by incorporating user feedback in the learning process and
improving computation speed. It merges adaboost, feature re-weighting
and relevance feedback into one framework and exploits the favorable
attributes of these methods. In the high level analysis, a novel
framework to develop a lexicon of high-level concepts with small
semantic gaps is constructed. By quantitatively studying and
formulating the semantic gap problem, these visually and semantically
consistent concepts are automatically selected and show their
promising application potential for concept detection, automatic
annotation, and multimedia information retrieval. Those algorithms are
successfully applied to several real-world applications on image
annotation, face recognition, and gene expression-based microarray
analysis.
Speaker Biography
Yijuan Lu is a Ph.D. candidate in the department of Computer Science
at University of Texas at San Antonio. Her research is in the areas of Multimedia
Information Retrieval, Machine Learning, and Bioinformatics.
Specifically, her primary research focuses on developing methodology
for semantic computing, multimodal fusions, subspace and manifold
learning, and Microarray Expression and Genomic Analysis (MEGA).
|
