Although 3D object scanning has come a long way in recent years, especially with the development of specialized hardware, automatic volumetric modeling of 3D objects still requires further development. Today, advanced hardware such as Microsoft's Hololens has revolutionized the marketplace because of meshes reconstruction capabilities that are assembled interactively. It can create 3D meshes of indoor scenes, with all the objects, floors and walls. Despite the technological advance both hardware and software, manual modeling is still the most common practice, making the market more expensive and requiring excessive production time. However, the advancement of new algorithms has opened new and precise ways for the automation of this process. This project aims to propose modern methods, based on algorithms such as SIFT-CNN, for object detachment from volumetric meshes with projection of textures on the mesh produced by Microsoft Hololens. The figure shows an example of a reconstruction. Image A has an object of interest, a wooden box. Figure B represents a mesh after the reconstruction of the environment with the wooden box, created with Hololens. Image C illustrates the application of a 3D filter on the mesh for smoothing, facilitating the process of separating the part of the mesh that represents the wooden box. The next step is the method for object detachment from the mesh, identifying objects and it's vertex then separating the connections with the mesh.