THESIS
2005
xii, 71 leaves : ill. ; 30 cm
Abstract
Scene understanding is one of the core subjects of computer vision. Given a single or a sequence of images only, a computer algorithm should be capable of analyzing pertinent data and translating them into meaningful entities for subsequent processing. In this thesis, we study the role of photometry in computer vision. Photometry is a class of measurement techniques that performs measurement to real world objects using images only. In the realm, of scene understanding, the measurement alone should provide sufficient information to allow high level visual processing....[
Read more ]
Scene understanding is one of the core subjects of computer vision. Given a single or a sequence of images only, a computer algorithm should be capable of analyzing pertinent data and translating them into meaningful entities for subsequent processing. In this thesis, we study the role of photometry in computer vision. Photometry is a class of measurement techniques that performs measurement to real world objects using images only. In the realm, of scene understanding, the measurement alone should provide sufficient information to allow high level visual processing.
To understand a scene from images only, we have to understand the components or fac-tors that explain the observation or measurement. In computer vision, it is known that surface reflectance gives the observed color of a pixel, and the surface orientation af-fects the reflectance, which is related to the underlying object geometry. Shadow is also an important cue for indicating the relationship among objects. This thesis consists of an in depth study on these interdependent three factors. Their respective measurements give rise to the proper computer vision algorithms for scene understanding: reflectance separation, photometric stereo and shadow extraction. In this thesis, we propose new and effective methods to tackle down the well-defined problems in each area, and we make contributions in robust scene understanding by exploiting photometric measure-ment and producing some of the best results to date.
Post a Comment