Unsupervised reconstruction of a Visual Hull in space, time and light domains
|Abstract:||This thesis presents an unsupervised image segmentation approach for obtaining a set of silhouettes along with the Visual Hull of an object observed from multiple viewpoints. The proposed approach can deal with mostly any type of appearance characteristics such as texture, similar background color, shininess, transparency besides other phenomena such as shadows and color bleeding. Compared to more classical methods for silhouette extraction from multiple views, for which certain assumptions are made on the object or scene, neither the background nor the object appearance properties are modeled. The only assumption is the constancy of the unknown background for a given camera viewpoint while the object is under motion. The principal idea of the approach is the estimation of the temporal evolution of each pixel over time which provides a stability measurement and leads to its associated background cost function. In order to cope with shadows and self-shadows, an object is captured under different lighting conditions. Furthermore, the information from the space, time and lighting domains is exploited and merged based on a Markov Random Field framework and the constructed energy function is minimized via graph-cut. Real-data experiments under different lighting condition show that the proposed approach allows for robust Visual Hull reconstruction of a variety of challenging objects such as objects made of shiny metal or glass.|
|Document Type:||Thèse de doctorat|
|Open Access Date:||23 April 2018|
|Collection:||Thèses et mémoires|
All documents in CorpusUL are protected by Copyright Act of Canada.