An iterative model for contrast detection, which accounts for the contrast invariance of orientation preference, has been recently suggested. The work here extends the iterative model by incorporating a cross-oriented suppression of simple cells in the primary visual cortex (V1). The modified model has a better performance in terms of robustness to noise, generates sharper edge responses while suppressing weak edges, and converges faster on equilibrium. The model exhibits a higher level of contrast invariance of orientation preference generating a clear pattern of edges in natural images.
展开▼
