Point Cloud Reference Frame Updates for Slice-Based Inter Prediction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional point cloud coding techniques face inefficiencies in computing resource utilization and errors in reference frame management due to the use of whole reference frames that are not spatially relevant and asynchronous updates in multi-slice frames.
Innovation Solution
The proposed methods involve updating the reference frame after encoding/decoding a current frame, deriving a reference point cloud sample before inter prediction, and conditionally signaling the attribute inter search range to improve coding efficiency and accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Area of stationary object
If whole reference frames are used for inter prediction, then the coverage area is complete, but the computing resource utilization becomes inefficient due to lack of spatial relevance
Solution Approach 1:
The patent divides the reference frame into multiple slices, where each slice corresponds to a specific spatial region. Instead of using the entire reference frame for prediction, only the relevant slice is utilized for inter prediction of the current frame, reducing unnecessary computation while maintaining complete spatial coverage where needed.
Solution Approach 2:
The patent applies different reference frame update strategies to different slices based on their spatial characteristics. Each slice can be independently updated and referenced, allowing the system to optimize computation by only processing locally relevant reference data rather than globally uniform processing.
2Ease of operation
If reference frames are updated asynchronously in multi-slice frames, then the processing can be distributed, but errors occur in reference frame management
Solution Approach 1:
The patent implements a feedback mechanism where the encoder and decoder both perform reference frame updates using the same slicing logic. The decoder receives slicing configuration information from the encoder and applies identical update rules, ensuring that both sides maintain consistent reference frames despite distributed asynchronous processing.
Solution Approach 2:
The patent creates a universal reference frame management system where the same slicing and update logic applies to all slices across different frames. This multi-functional approach ensures consistency throughout the system, allowing distributed processing while maintaining reliable reference frame management through unified rules.
3Device complexity
If conventional point cloud coding techniques are used, then the implementation is simple, but the coding efficiency is insufficient
Solution Approach 1:
The patent introduces dynamic reference frame updating where the reference frame is not fixed but continuously updated based on previously decoded frames. This dynamic approach improves coding efficiency by adapting to temporal changes in the point cloud data while maintaining manageable complexity through systematic update rules.
Solution Approach 2:
The patent performs reference frame updates after each frame decoding, preparing the reference frame for subsequent predictions in advance. This preliminary action ensures that when inter prediction is needed, the most up-to-date reference data is already available, improving coding efficiency without requiring complex real-time processing.
Data Source
AI summary
Embodiments of the present disclosure provide a method for point cloud coding. In the method, a conversion between a current frame of a point cloud sequence and a bitstream of the point cloud sequence is performed. A reference frame is updated based on the conversion. The reference frame is to be used for a further conversion between at least one subsequent frame of the point cloud sequence and the bitstream of the point cloud sequence.