To improve the generation of images on mobile devices for applications involving virtual and augmented reality, technologists have come up with a new approach for creating physical objects through the use of a point-and-shoot camera equipped with a flash. The new process was demonstrated using a Nikon D7000 camera plus a built-in camera of an Android mobile phone. The process is described as the capturing of spatially-varying bidirectional reflectance distribution functions. This process is key to obtaining an object’s real-world shape and appearance.
Current methods for capturing physical objects need special hardware setups, such as the use of laser scanners or with the use of multiple cameras. The new technique, as Laboratory Roots reports, achieves the same 3D rendering through the use of just a single camera. The new model uses an especially developed algorithm.
Commenting on the model, lead researcher Professor Min H. Kim stated: “To faithfully reproduce a real-world object in the VR/AR environment, we need to replicate the 3D geometry and appearance of the object.”
Kim added: “Traditionally, this has been either done manually by 3D artists, which is a labor-intensive task, or by using specialized, expensive hardware. Our method is straightforward, cheaper and efficient, and reproduces realistic 3D objects by just taking photos from a single camera with a built-in flash.”
The technology is described in greater detail in the following video:
The video shows examples of the novel approach being used to generate images of metal, wood, plastic, ceramic, and a finely detailed mini-statute of Nefertiti.
The new model has been presented in the article: “Practical SVBRDF Acquisition of 3D Objects with Unstructured Flash Photography.” The paper was presented at the conference SIGGRAPH Asia 2018.
In future work, the researchers hope to further simplify the capture process or extending the method to include dynamic geometry or larger scenes.