A novel hierarchical statistical model for count data modeling and its application in image classification

Authors:
Ali Shojaee Bakhtiari;Nizar Bouguila
Affiliations:
Department of Electrical and Computer Engineering and Concordia Institute for Information Systems Engineering, Concordia University, Montreal, QC, Canada;Department of Electrical and Computer Engineering and Concordia Institute for Information Systems Engineering, Concordia University, Montreal, QC, Canada
Venue:
ICONIP'12 Proceedings of the 19th international conference on Neural Information Processing - Volume Part II
Year:
2012

Citing 10
Cited 1

Object Recognition from Local Scale-Invariant Features

ICCV '99 Proceedings of the International Conference on Computer Vision-Volume 2 - Volume 2
A Bayesian Hierarchical Model for Learning Natural Scene Categories

CVPR '05 Proceedings of the 2005 IEEE Computer Society Conference on Computer Vision and Pattern Recognition (CVPR'05) - Volume 2 - Volume 02
MML-Based Approach for High-Dimensional Unsupervised Learning Using the Generalized Dirichlet Mixture

CVPR '05 Proceedings of the 2005 IEEE Computer Society Conference on Computer Vision and Pattern Recognition (CVPR'05) - Workshops - Volume 03
Hierarchical Dirichlet model for document classification

ICML '05 Proceedings of the 22nd international conference on Machine learning
Generalized liouville distribution

Computers & Mathematics with Applications
A Liouville-based approach for discrete data categorization

RSFDGrC'11 Proceedings of the 13th international conference on Rough sets, fuzzy sets, data mining and granular computing
An Expandable Hierarchical Statistical Framework for Count Data Modeling and Its Application to Object Classification

ICTAI '11 Proceedings of the 2011 IEEE 23rd International Conference on Tools with Artificial Intelligence
Image retrieval by examples

IEEE Transactions on Multimedia
Medical Image Categorization and Retrieval for PACS Using the GMM-KL Framework

IEEE Transactions on Information Technology in Biomedicine
Support vector machines for histogram-based image classification

IEEE Transactions on Neural Networks

Object clustering and recognition using multi-finite mixtures for semantic classes and hierarchy modeling

Expert Systems with Applications: An International Journal

Quantified Score

Hi-index	0.00

Visualization

Abstract

The problem that we elaborate in this work is developing and comparing statistical models for learning hierarchical image categories from a structural point of view. Previously different statistical models have been proposed based on different statistical schemes for dealing with hierarchical structures. In this work following the lead of the previous models we develop our own hierarchical model and we make a thorough comparison between the existing and the proposed models. Our main contribution in this work is the utilization of Beta-Liouville distribution as a replacement for Dirichlet distribution, which is traditionally used for prior distribution modeling, and deriving the criteria for making it compatible to hierarchical data modeling. For the development of our statistical model, we make extensive use of the Bag of the visual words model and the concept of count data in machine learning.