Skip to Main content Skip to Navigation
Theses

Reconstruction de scène dynamique à partir de plusieurs vidéos mono et multiscopiques par hybridation de méthodes « silhouettes » et « multi-stéréovision »

Abstract : Accurate reconstruction of a 3D scene from multiple cameras offers 3D synthetic content to be used in many applications such as entertainment, TV, and cinema production. This thesis is placed in the context of the RECOVER3D collaborative project, which aims is to provide efficient and quality innovative solutions to 3D acquisition of actors. The RECOVER3D acquisition system is composed of several tens of synchronized cameras scattered around the observed scene within a chromakey studio in order to build the visual hull, with several groups laid as multiscopic units dedicated to multi-baseline stereovision. A multiscopic unit is defined as a set of aligned and evenly distributed cameras. This thesis proposes a novel framework for multi-view 3D reconstruction relying on both multi-baseline stereovision and visual hull. This method’s inputs are a visual hull and several sets of multi-baseline views. For each such view set, a multi-baseline stereovision method yields a surface which is used to carve the visual hull. Carved visual hulls from different view sets are then fused iteratively to deliver the intended 3D model. Furthermore, we propose a framework for multi-baseline stereo-vision which provides upon the Disparity Space (DS), a materiality map expressing the probability for 3D sample points to lie on a visible surface. The results confirm i) the efficient of using the materiality map to deal with commonly occurring problems in multi- baseline stereovision in particular for semi or partially occluded regions, ii) the benefit of merging visual hull and multi-baseline stereovision methods to produce 3D objects models with high precision.
Complete list of metadata

Cited literature [89 references]  Display  Hide  Download

https://hal.archives-ouvertes.fr/tel-02883201
Contributor : Nicolas Passat <>
Submitted on : Sunday, June 28, 2020 - 8:15:49 PM
Last modification on : Monday, May 17, 2021 - 2:52:12 PM

File

45169_ISMAEL_2016_archivage.pd...
Files produced by the author(s)

Identifiers

  • HAL Id : tel-02883201, version 1

Collections

Citation

Muhannad Ismael. Reconstruction de scène dynamique à partir de plusieurs vidéos mono et multiscopiques par hybridation de méthodes « silhouettes » et « multi-stéréovision ». Vision par ordinateur et reconnaissance de formes [cs.CV]. Université de Reims Champagne-Ardenne, 2016. Français. ⟨tel-02883201⟩

Share

Metrics

Record views

37

Files downloads

83