Active/space-variant object recognition

Abstract The problem of object recognition is addressed. In the literature this task has been generally considered in a passive perspective, where everything is static and there is no definite relation between the object and its environment. In this paper, several aspects related to the application of active vision techniques to object recognition are discussed. The capability of the observer to move is very important to give a better description of the object during acquisition of the model database, and also for recognition. The recognition task can be simplified considerably by defining classes of expected objects based on contextual information. Moreover, a selective attentional mechanism allows us to reduce the amount of information needed to describe a database of objects. This is accomplished both at the task level, by performing planned fixations, and at the sensor level, by adopting a space-variant sampling of the image. The face recognition problem based on the face-space approach is considered to demonstrate the advantage of adopting an active retina in recognition tasks. By using an active space-variant retina, the size of the database is considerably reduced and, consequently, so too is the processing time for recognition.