Modeling the perception of color patterns with HMAX.
Project Leaders: Julien RENOULT, François MOLINO & Frédéric GENIET
HMAX is an artificial neural network that models perception of color patterns in Humans. HMAX has been developed at Maximilian Reisenhuber Lab. HMAX is also a powerful tool to model the perceived similarity between entire visual scenes or phenotypes.We propose a fast MATLAB implementation of HMAX that can be easily adapted to the color vision of other Vertebrate species. Our computer scientist Bastien GUYL also developed sparse-HMAX, which optimizes sparseness in the S2 layer. Notably, sparse-HMAX can be used to analyse how efficiently processed a stimulus is.