The Caltech component of the proposed Center, """"""""The Role of Attention in Detecting and Recognizing Objects,"""""""" is organized around the central theme of attentional aspects of object recognition, using techniques that the Koch lab has employed successfully in the past: human psychophysics, computational (abstract) and neuronally-plausible (biophysical) modelling. The research is organized in 3 aims, each of them with obvious relevance to different groups in the center:
AIM 1 : Human psychophysics of attention and recognition in natural scenes will parallel electrophysiological work in monkeys (DiCarlo) and determine the effects of """"""""physical"""""""" distance between stimuli (i.e. clutter) and """"""""similarity"""""""" distance between targets and distractors (i.e. task complexity) on a task's attentional requirements. This will help determine the limits of the current HMAX feed-forward recognition system (Riesenhuber and Poggio).
AIM 2 : In light of these constraints, the computational model of saliency previously developed in Koch's group will be integrated with the HMAX feed-forward recognition system (Riesenhuber and Poggio) to implement attentional modulation of object recognition.
AIM 3 : Finally, the Koch group will make use of its experience in the field of biophysical modelling to study detailed, biologically plausible models of the MAX operation at the level of simple neuronal feedback circuits or individual cortical neurons which will aim to account for electrophysiological results obtained in the Ferster group.
Showing the most recent 10 out of 12 publications
