ISOBMFF Point Cloud Metadata for Selective Attribute Decoding
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for compressing and storing point clouds with no time information, such as map data, are inefficient as they require decoding unnecessary attribute information, leading to processing inefficiencies.
Innovation Solution
A method for generating metadata that includes reproduction information and feasibility information for bitstreams generated from 3D data with no time information, allowing efficient storage and reproduction of point clouds in a file structure using ISOBMFF technology, such as HEIF, by separating and signaling necessary attributes and geometry.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If conventional point cloud compression methods (V-PCC or G-PCC) are used to encode and store point clouds with no time information, then the point cloud data can be stored in a file structure, but all attribute information must be decoded even when not needed, leading to processing inefficiency
Solution Approach 1:
The patent segments the point cloud data into multiple independent bitstreams, each representing a specific attribute (e.g., geometry, color, reflectance). This segmentation allows the reproduction device to selectively decode only the necessary attribute bitstreams based on reproduction feasibility information, avoiding the need to decode all attributes as required by conventional methods.
Solution Approach 2:
The patent extracts and separates the necessity information for each attribute into metadata. This extraction enables the reproduction device to identify which attributes are actually needed for reproduction and skip decoding of unnecessary attributes, directly addressing the processing inefficiency caused by conventional methods that require decoding of all attributes.
2Reliability
If conventional point cloud compression methods are used, then complete point cloud data can be stored, but the file structure lacks information to determine which attributes are necessary for reproduction, causing computational overhead
Solution Approach 1:
The patent performs preliminary action by embedding reproduction feasibility information into the file structure during the encoding phase. This information indicates which attributes are necessary for reproduction, allowing the reproduction device to make decisions without complex computational analysis, thereby reducing device complexity while maintaining data completeness.
3Manufacturing precision
If all attribute information is decoded in conventional methods, then complete reproduction is possible, but processing time and computational resources are wasted on unnecessary attributes
Solution Approach 1:
The patent extracts reproduction feasibility information from the encoded data and stores it in the file structure. This extraction enables the reproduction device to identify and decode only the necessary attribute bitstreams, reducing computational energy consumption while ensuring accurate reproduction of the required point cloud attributes.
Data Source
AI summary
There is provided an information processing device and an information processing method by which reproduction of a point cloud having no time information can be handled. Metadata that includes reproduction information necessary to reproduce encoded data obtained by encoding one frame of a point cloud by V-PCC or G-PCC and that includes reproduction feasibility information to be used to determine feasibility of reproducing the encoded data is generated. The encoded data and the metadata are stored in a file having a predetermined file structure using the ISOBMFF technology. The present technology is applicable to a data generation device that generates a file into which encoded data of a point cloud having no time information is stored, for example.