When we are presented with an image our visual system does not treat all the information in the image equally. We quickly focus on restricted subsets of the image. This action is known as focus of attention. For example, when driving a car road signs, bicyclists, and other cars quickly catch our attention and we tend to ignore the sky, trees and other details in the visual image.
In this seminar we report on an investigation of a computer vision system model that incorporates the characteristics of the early stages of the human visual system in an attempt to duplicate our visual focus of attention. The computer vision system uses color difference maps, edge maps and intensity difference maps at different spatial resolutions as features. The results of the individual maps are then combined to form a focus of attention mechanism for an image.
In Part I of this seminar I discussed a classical method to solve computer vision recognition problems and gave a brief introduction into a biologically based model of vision. In Part II I will briefly review the material presented in the first seminar and continue to develop the biologically inspired model. This computer vision system uses color difference maps, edge maps and intensity difference maps at different spatial resolutions as features. The results of the individual maps are then combined to form a focus of attention mechanism for an image.
Colloquia Series page.