Passive 3D reconstruction (static scenes)
The project aims to automatically infer the shape of a static object from a set of pictures. Recent efforts attempt to combine together information of different passive methods. Critical issues in this research are the choice of data and how to combine such data in order to increase the overall information. The combination of stereo matching and silhouette information has recently received considerable attention both for obtaining high quality 3D models and for modelling 3D dynamic scenes. In this project we aim at building a 3D shape recovery system which fuse together silhouette, texture and shadow information [Ballan2006].The reconstruction pipeline is the following:
Video: [link]
Video: [link]
Once the shape is estimated, its texture is recovered using a pyramidal blending technique based on wavelets [Brusco2005].
Results:
Frog model [WRL]
(silhouette + stereo)
Cube model [WRL]
(silhouette + stereo + shadow)
Statue model [WRL]
(silhouette + stereo + shadow)
|
Human models (silhouette + stereo) Mesh: 600k faces Texture: 6000x3500 pixels Model 1: [OBJ] [MTL] [TEX] Model 2: [OBJ] [MTL] [TEX] Model 3: [OBJ] [MTL] [TEX] |
|||
Publications:
Multimodal 3D Shape Recovery from Texture, Silhouette and Shadow InformationL. Ballan and G. M. Cortelazzo [PDF] [bibtex]3DPVT 2006, Chapel Hill, USA |
Efficient 3D Modeling Of Sculptures From PhotographsL. Ballan, N. Brusco and G. M. Cortelazzo [bibtex]EVA 2006 (The Foremost European Electronic Imaging Events in the Visual Arts), Florence, Italy |
3D Passive Shape Recovery from Texture and Silhouette InformationL. Ballan, N. Brusco and G. M. Cortelazzo [PDF] [bibtex]CVMP 2005 (European Conference on Visual Media Production), London, UK |
Passive Reconstruction of High Quality Textured 3D Models of Works of ArtN. Brusco, L. Ballan and G. M. Cortelazzo [PDF] [bibtex]VAST 2005 (International Symposium on Virtual Reality, Archeology and Cultural Heritage), Pisa, Italy |
Applicazione dei Modelli Deformabili alla Digitalizzazione 3D PassivaL. Ballan [PDF] [bibtex] (italian)Master's thesis, Department of Information Engineering, University of Padova, 2005 |
Passive 3D reconstruction (dynamic scenes)
Dynamic scene modeling is a challenging problem in computer vision. Many techniques have been developed in the past to address such a problem but most of them focus on achieving accurate reconstructions in controlled environments, where the background and the lighting are known and the cameras are fixed and calibrated. Recent approaches have relaxed these requirements by applying these techniques to outdoor scenarios. The problem however becomes even harder when the cameras are allowed to move during the recording since no background color model can be easily inferred. In this paper we propose a new approach to model dynamic scenes captured in outdoor environments with moving cameras. A probabilistic framework is proposed to deal with such a scenario and to provide a volumetric reconstruction of all the dynamic elements of the scene.Publications:
Modeling Dynamic Scenes Recorded with Freely Moving CamerasA. Taneja, L. Ballan and M. Pollefeys [PDF] [bibtex]ACCV 2010, Queenstown, New Zealand |
