A media type conversion framework between 3D object and MIV

Abstract

This paper proposes a framework for converting 3D content into MPEG Immersive Video (MIV)-compliant formats. Unlike existing MIV entity concepts, which are limited to predefined scenarios, the proposed approach supports heterogeneous objects and enables flexible composition on the encoder side and dynamic rendering on the decoder side. The framework accommodates diverse camera projections, spatial origins, and scaling requirements, allowing 3D objects with color and spatial data to be transformed into MIV-compatible entities. Furthermore, it incorporates MIV coding verification for the generated heterogeneous objects to ensure accuracy and reliability.