*Project Title*
Data Fusion Algorithms for the OntoSAIA System

*Project Description* 

OntoSAIA is a tool that helps image retrieval and annotation, integrating the approaches based on image content, 
keywords, and ontologies. Using OntoSAIA, it is possible to retrieve, visualize, add, remove, or annotate images 
of a collection, as well as explore the associated ontologies. For OntoSAIA, each image is associated with an 
annotation that is composed of terms, and with a set of feature vectors for the different image descriptors. 

The image annotation process is based on term suggestion which takes into consideration two main points: (i) the 
annotation of a new image is strongly related to previous annotations; and (ii) the annotation of an image is related 
to the content of the image. The suggestion of terms in the latter case takes into account the similarity between the 
image to be annotated and the images already present in the image database. If two images have similar content, it 
is possible that both depict similar objects and, consequently, should have similar annotations.  Therefore, after the 
user selects the image to be annotated, it is compared, according to the image descriptor (implicitly or explicitly) 
selected by the user, to the images present in the image database. Afterwards, the annotations of the most similar 
images are analyzed and the terms that are most frequent in these annotations are suggested. 
The objective of this project is the specification and implementation of data fusion strategies to be incorporated into 
the suggestion engine of the OntoSAIA tool. These fusion strategies will be used to combine suggestions based on 
ontologies, image-content information, and database term frequency.

*Language/technology skills*
Java

*Number of People*
1-2

*Project Resources*

1. R. B. Freitas and R. da S. Torres, OntoSAIA: Um ambiente Baseado em Ontologias para Recuperação e Anotação 
Semi-Automática de Imagens , Primeiro Workshop de Bibliotecas Digitais, 60-79, Uberlandia, Simpósio Brasileiro de 
Banco de Dados, 2005.
http://www.ic.unicamp.br/~rtorres/freitas05wdl.pdf