Omnidirectional Video Metadata Mapping for Planar Transmission
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
There is no standardized file format for transmitting and receiving omnidirectional media, such as Virtual Reality (VR) content, which hinders the efficient storage, exchange, management, and presentation of immersive 360-degree content.
Innovation Solution
A method and apparatus for transmitting and receiving omnidirectional media using metadata that maps omnidirectional video content onto planar video, allowing for reconstruction and rendering on various geometries like spheres, cylinders, and platonic solids, utilizing the ISO Base Media File Format (ISO BMFF) and Omnidirectional Media Application Format (OMAF).
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Manufacturing precision
If omnidirectional video is transmitted directly without mapping to planar video, then the original 360-degree view quality is preserved, but the transmission and storage efficiency deteriorates due to lack of standardized format
Solution Approach 1:
The patent transforms omnidirectional video from its native 360-degree spherical coordinate system into a 2D planar video format using standardized projection methods (equirectangular, cubic, etc.). This dimensional transformation enables the video to be stored and transmitted using conventional 2D video codecs and file formats, significantly improving transmission and storage efficiency while maintaining the ability to reconstruct the original omnidirectional view at the receiving end through metadata-guided rendering
2Productivity
If omnidirectional video is mapped to planar video for transmission, then transmission efficiency is improved, but the complexity of reconstruction and rendering increases
Solution Approach 1:
The patent performs preliminary actions by embedding comprehensive metadata (projection type, resolution, geometry information, region mappings) into the planar video container during the encoding phase. This pre-packaging of all necessary reconstruction information eliminates the need for complex real-time calculations at the receiving end, as the decoder can directly use the embedded metadata to reconstruct the omnidirectional video without iterative optimization or complex geometric computations
Solution Approach 2:
The patent introduces metadata as an intermediary layer between the planar video and the omnidirectional reconstruction process. This metadata acts as a bridge that carries all necessary geometric and projection information, allowing the receiving device to automatically reconstruct the omnidirectional view without requiring complex algorithms. The metadata includes projection type indicators, resolution parameters, and region-mapping information that guide the reconstruction process
3Productivity
If a standardized file format is implemented for omnidirectional media, then storage and management efficiency is improved, but the flexibility in handling various geometries (sphere, cylinder, platonic solids) may be reduced
Solution Approach 1:
The patent creates a universal file format (OMAF based on ISO BMFF) that can handle multiple geometry types (sphere, cylinder, platonic solids) through a unified metadata structure. The format includes geometry type identifiers and parameter fields that can accommodate different geometric configurations, allowing a single standardized format to serve multiple purposes and geometry types without requiring separate proprietary formats for each geometry
Data Source
Figure 1A
Figure 1B
Figure 1C~2
AI summary
The present disclosure relates to a method for transmitting metadata of an omnidirectional image by a transmission apparatus, the method comprising the steps of: projecting an omnidirectional image into a planar image based on a 3-dimensional (3D) geometry corresponding to the omnidirectional image; and transmitting metadata including type information of the 3D geometry, information for a region on the projected planar image, and information for an area on the omnidirectional image which is mapped to the region.