A spatially aware generative model for image classification, topic discovery and segmentation

Authors:
Iván González-Díaz;Darío García-García;Fernando Díaz-de-María
Affiliations:
Department of Signal Theory and Communications, Universidad Carlos III, Leganés, Madrid, Spain;Department of Signal Theory and Communications, Universidad Carlos III, Leganés, Madrid, Spain;Department of Signal Theory and Communications, Universidad Carlos III, Leganés, Madrid, Spain
Venue:
ICIP'09 Proceedings of the 16th IEEE international conference on Image processing
Year:
2009

Citing 7
Cited 1

Unsupervised learning by probabilistic latent semantic analysis

Machine Learning
Latent dirichlet allocation

The Journal of Machine Learning Research
Distinctive Image Features from Scale-Invariant Keypoints

International Journal of Computer Vision
Discovering Objects and their Localization in Images

ICCV '05 Proceedings of the Tenth IEEE International Conference on Computer Vision (ICCV'05) Volume 1 - Volume 01
A Comparison of Affine Region Detectors

International Journal of Computer Vision
Semantic-Shift for Unsupervised Object Detection

CVPRW '06 Proceedings of the 2006 Conference on Computer Vision and Pattern Recognition Workshop
A fast recursive shortest spanning tree for image segmentation and edge detection

IEEE Transactions on Image Processing

A region-centered topic model for object discovery and category-based image segmentation

Pattern Recognition

Quantified Score

Hi-index	0.01

Visualization

Abstract

For the last few years bag-of-words models have been succesfully applied to the information retrieval field. However their application to visual content suffers from an important shortcoming: they model images as sets of unordered visual words rather than consider their spatial and geometric layout. Visual information is highly organized along the dimensions of an image and algorithms should make use of this to enhance the performance of several visual processing tasks. In this paper, a generative model is proposed that fuses both the local information obtained from visual words and the global geometric layout given by a previous segmentation of the image. Furthermore, the model considers inter-region influences so topics can spread along the image and, thus, generate final segmentations in which regions represent semantic concepts. The proposed model is succesfully tested on three different tasks.