AR/VR Geometry Conversion for Compact Multi-Domain Bitstreams
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current AR/VR systems inefficiently represent and transmit geometry data due to redundant formats and formats not optimized for small bitstream sizes, leading to increased data redundancy and inefficiency.
Innovation Solution
A unified, compact representation of geometry data is achieved through a converter that transforms spatial object representations between different domains, using a geometry data conversion block and additional metadata to ensure consistent data formats across signal processing blocks, reducing redundant data transmission.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If geometry data is transmitted multiple times in different formats to meet input parameter requirements of different signal processing blocks, then each block can process data in its preferred format, but the bitstream size increases and data redundancy increases
Solution Approach 1:
The geometry data is segmented into different representation domains (e.g., mesh domain, primitive domain, edge domain) and each domain is optimized for specific signal processing blocks. The converter divides the conversion task into domain-specific transformations, allowing each block to receive data in its preferred format without requiring all data to be transmitted in all formats.
Solution Approach 2:
A geometry data converter acts as an intermediary between the unified geometry data source and the different signal processing blocks. This converter transforms the unified geometry data into domain-specific representations (mesh, primitive, edge) that match the input requirements of early reflection, diffraction, and other processing blocks, eliminating the need to transmit multiple copies of the same data.
2Adaptability or versatility
If multiple geometry representations are transmitted individually through meshes and primitives payloads, then each signal processing block receives data in its required format, but the representation becomes highly redundant and inefficient
Solution Approach 1:
The system introduces a unified geometry data structure that serves multiple functions by representing spatial objects in a single, standardized format. This universal representation can then be converted to different domain-specific formats (mesh, primitive, edge) as needed by different signal processing blocks, eliminating the need for separate transmission of multiple redundant representations.
Solution Approach 2:
The patent adds a new dimension to geometry data representation by introducing a unified domain that can be transformed into multiple specialized domains. Instead of transmitting data in multiple formats simultaneously, the system transmits data in one unified format and performs conversions at the receiving end, adding the dimension of format conversion to the data transmission model.
3Productivity
If a binary format is used for transmitting diffraction payload data, then data can be transmitted efficiently, but the format is not optimized for small bitstream sizes
Solution Approach 1:
The patent changes the parameters of the binary format by introducing domain-specific conversion parameters and metadata that enable efficient encoding of geometry data in the unified domain. This allows the binary format to be optimized for small bitstream sizes while maintaining transmission efficiency, as the converter can process the compact data and generate appropriate domain-specific representations without requiring large bitstreams.
Data Source
AI summary
An apparatus according to an embodiment. The apparatus comprises a receiving interface for receiving a plurality of spatial object representations of a plurality of spatial objects, wherein each of the plurality of spatial object representations is represented in at least one of two or more spatial representation domains. Moreover, the apparatus comprises a converter for converting one or more of the plurality of spatial object representations from a first one of the two or more spatial representation domains into a second one of the two or more spatial representation domains.