Contours Inhibition for Object Tracking: a Preliminary Study

Overview & Synopsis

This study introduces an innovative approach to object tracking by integrating the concept of contour inhibition, a key feature of the human visual system, into computational models. The human ability to discern objects from complex backgrounds by focusing on object edges, rather than textures, provides the inspiration for this work. By emulating this selective attention mechanism, the aim is to include this edge detection methodology for improved object tracking. Traditional tracking methods rely on identifying features critical for tracing objects within video sequences. The proposed approach aims to utilize selected contour pixels and their associated direction as the primary features. The initial results from this study demonstrate the potential of considering contour inhibition for tracking through the direction of each preserved contour pixel.