This paper presents the approach used by the University of Alicante in the ImageCLEF 2004 adhoc retrieval task. This task is performed by multilingual search requests (topics) against an historic photographic collection in which images are accompanied with English captions. This approach uses these captions to make the retrieval and is based in a set constraints and preferences that will allow the rejection or the scoring of the images for the retrieval task. The constraints are implemented through a set of co-occurrence patterns based on regular expressions and the approach is extended in one of the experiments with the use of WordNet synonym relations.