Title :
Supervised topic model for automatic image annotation
Author :
Putthividhya, D. ; Attias, H.T. ; Nagarajan, S.S.
Author_Institution :
Inst. for Neural Comput., UCSD, La Jolla, CA, USA
Abstract :
This paper presents a new probabilistic model for the task of image annotation. Our model, which we call sLDA-bin, extends supervised Latent Dirichlet Allocation (sLDA) model to handle a multi-variate binary response variable of the annotation data. Unlike correspondence LDA (cLDA), the association model in sLDA allows each caption word to be associated with more than 1 image region and is thus more appropriate for annotation words that globally describe the scene. By modeling the response variable as a multi-variate Bernoulli, we introduce a tight convex variational bound for the logistic function and derive an efficient variational inference algorithm based on mean-field approximation. Our model compares favorably with cLDA on an image annotation task, as demonstrated by a superior caption prediction probability.
Keywords :
image retrieval; probability; automatic image annotation; mean-field approximation; multivariate Bernoulli; multivariate binary response variable; probabilistic model; sLDA-bin; supervised latent dirichlet allocation model; supervised topic model; tight convex variational bound; variational inference algorithm; Approximation algorithms; Image retrieval; Inference algorithms; Layout; Linear discriminant analysis; Logistics; Probability distribution; Radiology; Signal processing algorithms; Vocabulary; Automatic Image Annotation; Image Retrieval; Multimedia Signal Processing; Probabilistic Graphical Models; Statistical Topic Models;
Conference_Titel :
Acoustics Speech and Signal Processing (ICASSP), 2010 IEEE International Conference on
Conference_Location :
Dallas, TX
Print_ISBN :
978-1-4244-4295-9
Electronic_ISBN :
1520-6149
DOI :
10.1109/ICASSP.2010.5495341