Best of Both Worlds: Merging 360˚ Image Capture with 3D Reconstructed Environments for Improved Immersion in Virtual Reality

Public Deposited

Analytics

Last Modified

Creator

Rajkumar, Vijay
- Affiliation: College of Arts and Sciences, Department of Computer Science

Abstract

With the recent proliferation of high-quality 360° photos and video, consumers of virtual reality (VR) media have come to expect photorealistic immersive content. Most 360° VR content, however, is captured with monoscopic camera rigs and inherently fails to provide users with a sense of 3D depth and 6 degree-of-freedom (DOF) mobility. As a result, the medium is significantly limited in its immersive quality. This thesis aims to demonstrate how content creators can further bridge the gap between 360° content and fully immersive real-world VR simulations. We attempt to design a method that combines monoscopic 360° image capture with 3D reconstruction -- taking advantage of the best qualities of both technologies while only using consumer-grade equipment. By mapping the texture from panoramic 360° images to the 3D geometry of a scene, this system significantly improves the photo-realism of 3D reconstructed spaces at specific points of interest in a virtual environment. The technical hurdles faced during the course of this research work, and areas of further work needed to perfect the system, are discussed in detail. Once perfected, a user of the system should be able to simultaneously appreciate visual detail in 360-degrees while experiencing full mobility, i.e., to move around within the immersed scene.

Date of publication

Keyword

DOI

Resource type

Rights statement

Note

Advisor

Degree

Academic concentration

Honors level

Degree granting institution

Graduation year

Language

Relations

Thumbnail	Title	Date Uploaded	Visibility	Actions
	2017_Rajkumar.pdf	2019-05-17	Public	Download