Abstract
With the increase of data exchange and latest technological and social developments, multimedia contents are becoming an important part of global trafic. Today, 3D objects are used in a large number of applications, for example, medical applications, simulations, video games, animation and special effects. 3D object usage by the general public has become a lucrative market that can take the form of 3D object downloading platforms with various 3D formats.This thesis, in collaboration with the company STRATEGIES, concerns the 3D object protection, and more particularly 3D meshes against fradulent and illegal uses. These 3D meshes represent surface models of shoes and leather goods produced by customers using digital solutions proposed by STRATEGIES. First, we propose a new method to insert secret data much more efficiently in terms of execution time on very large meshes than the previous method developed in collaboration with the company STRATEGIES. We are also exploring selective encryption approaches to control access to very high quality content according to user needs. In this context, we propose to use selective encryption approaches on the geometric data of 3D objects in order to protect the visual content of these objects according to different use cases and different data representations.In a second research axis, we study the application of secret sharing methods to the domain of 3D objects. Secret sharing is an approach that seeks to divide secret content between multiple users and allows certain subgroups of users to reconstruct the secret. Secret sharing is a redundancy system that allows you to reconstruct the secret even if some users have lost their information. Secret 3D object sharing is a poorly researched domain used to protect a 3D object between collaborators. We propose new secret 3D object sharing methods using selective encryption approaches and providing hierarchical properties where users have different access rights to 3D content based on their position in a hierarchical structure.Finally, the third research axis developed in this thesis deals with the analysis of the visual confidentiality of 3D objects selectively encrypted more or less strongly. Indeed, depending on the scenario, our 3D selective encryption methods provide results that can be more or less recognizable by users. However, the metrics used to evaluate the quality of 3D objects do not distinguish two selectively encrypted 3D objects with different levels of confidentiality. So, we present the construction of a databse of selectively encrypted 3D objects in order to realize subjective assessments of visual confidentiality and try to build a new metric correlated with evaluations obtained by the human visual system.