Abstract

If moving objects in low-resolution 2D video imagery are placed in their 3D context, object uncertainties can be removed. We consider the case of detecting distant humans subtending only tens of pixels in digital video.

© 2008 Optical Society of America

PDF Article