A neuro-cognitive visual system for object recognition based on testing ofinteractive attentional top-down hypotheses

We propose an extension of a systemic model for object recognition formulat ed by Rybak et al (1998 Vision Research 38 2387-2400) which is based on the functional organisation of the visual systems in primate brains. In contra st to the learning and recognition scheme of Rybak et al we do not assume a behavioural paradigm, ie a visuomotor programmed scanpath that determines the sequence of foveation on the different parts of the object. As in the b asis architecture of Rybak et al, the system modules are separated into 'wh at'-like subsystems corresponding to the ventral occipito-inferotemporal vi sual path and 'where'-like complexes analogous to the dorsal occipito-parie tal visual path. The 'what' system analyses local features in the actual fo veation as in Rybak et al. But, in our case, the 'where' memory, instead of programming a behavioural scanpath, scores the spatial relationship betwee n successive fixation and the spatial relationship between the associated m ain edges. The recognition is based on the identification of parts and thei r spatial relationship. This gives the learning and recognition mechanisms more flexibility in the sense that, for recognising an object, several diff erent fixation sequences may be accepted.