Vous êtes ici : GIPSA-lab > Animation > Séminaires

Séminaire du département Images et Signal du 13/06/2013 à 15h00


The Role of V4 During Natural Vision

Intervenant : Julien Mairal, Inriaples - équipe LEAR

Lieu : salle Mont-Blanc


Résumé :

The functional organization of area V4 in the mammalian ventral visual pathway

is far from being well understood. V4 is believed to play an important role in

the recognition of shapes and objects and in visual attention, but its

complexity makes it hard to analyze. Individual cells in V4 have been shown to

exhibit a large diversity of preferences to visual stimuli characteristics,

including orientation, curvature, motion, color and texture.  Such observations

were for a large part obtained from electrophysiological and imaging studies,

when a subject (monkey or human) is shown a sequence of artificial stimuli

during data acquisition. In our study, we intend to go beyond such an approach

and analyze a population of V4 neurons in naturalistic conditions. More

precisely, we record responses from V4 neurons to grayscale still natural

images---that is, discarding color and motion content. We propose a new

computational model for V4 that does not rely on any pre-defined image features

but only on invariance and sparse coding principles. Our approach is the first

to achieve comparable prediction performance for V4 as for V1 cells on

responses to natural images. Our model is also interpretable using sparse

principal component analysis. In the neuron population observed and based on

our computational model, we discover as our main finding two groups of neurons:

those selective to texture versus those selective to contours. This supports

the thesis that one primary role of V4 is to extract objects from background in

the visual field.  Moreover, our study also confirms the diversity of V4

neurons. Among those selective to contours, some of them are selective to

orientation, others to acute curvature features.


This is a joint work with Yuval Benjamini, Ben Willmore, Michael Oliver, Jack

Gallant and Bin Yu. This work was performed at UC Berkeley.

GIPSA-lab, 11 rue des Mathématiques, Grenoble Campus BP46, F-38402 SAINT MARTIN D'HERES CEDEX - 33 (0)4 76 82 71 31