This article presents a new perspective to image retrieval based on multivariate factor analysis. Modern imaging modalities provide an overwhelming amount of information that cannot be appropriately handled without computerized tools. Image retrieval methods aim to retrieve relevant images from a collected database, based on their content. The problem is made more difficult due to the high dimension of the variable space in which the images are represented. Factor analysis is able to minimize data redundancy and reveal hidden patterns. The characterization of complex features such as shape can be performed in a new lower-dimensional basis, in which the variables account for the correlation among regions of interest. The similarity between images is computed based on a set of vector variables obtained from image registration. Relevance feedback is used to iteratively determine the importance of each factor in the similarity function. The factors that are supposed to encompass the user?s preferences are implicitly detected.