A model of surface detection and orientation tuning in primate visual cortex

In this contribution we propose that surface detectors, responding to areas of homogeneous gray-level may be used to support orientation selective cells in V1. By inhibiting orientation detectors within their input region, surface detectors act as local feedback system, emphasizing objects with large, continuous contours and enable the fast generation of an initial hypothesis about the input. The performance of this model is demonstrated, using a set of real-world images. We demonstrate that for images with a high degree of low-frequency clutter, the performance of orientation detection can be considerably improved.