3D Point Cloud Encoding and Decoding With Same-Layer Reference
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing three-dimensional data encoding and decoding methods lack efficiency in compressing and transmitting large volumes of point cloud data, particularly in systems using multiple codecs and formats.
Innovation Solution
Classify three-dimensional points into layers based on geometry information and allow same-layer reference for attribute information encoding and decoding, generating a bitstream that includes encoded attribute information.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Quantity of substance
If point cloud data is compressed using existing encoding methods, then data volume is reduced for accumulation and transmission, but coding efficiency is insufficient
Solution Approach 1:
The patent segments point cloud data into multiple layers based on geometric information (depth, distance from camera). This segmentation enables selective processing and reference mechanisms that improve compression efficiency. By dividing the data into layers, the system can apply different encoding strategies to different parts of the point cloud, reducing overall data volume while maintaining coding efficiency through optimized layer-based processing.
Solution Approach 2:
The patent performs preliminary classification of points into layers before encoding based on geometric attributes. This preliminary organization enables the encoding system to prepare reference structures in advance, allowing more efficient prediction and compression during the actual encoding process. The preliminary layer classification facilitates better prediction models that reduce the information needed to represent each point.
2Device complexity
If attribute information is encoded without same-layer reference, then encoding process is simpler, but coding efficiency is lower
Solution Approach 1:
The patent implements a feedback mechanism where encoded attribute information from same-layer points is used to generate prediction values for current points. This feedback loop allows the system to continuously improve encoding efficiency by using already-processed information from the same layer to predict and compress current data, reducing the information entropy and improving compression ratios without excessive complexity increase.
Solution Approach 2:
The patent changes the encoding parameters by introducing same-layer reference mechanisms that modify how attribute information is predicted and encoded. By allowing references to points within the same layer, the system changes the prediction model parameters to better capture local spatial relationships, thereby improving coding efficiency while managing complexity through controlled parameter modifications.
Data Source
AI summary
A three-dimensional data encoding method includes: encoding attribute information of three-dimensional points, the three-dimensional points being classified into one or more layers, based on geometry information of the three-dimensional points; and generating a bitstream including the attribute information encoded. In encoding of attribute information of a current three-dimensional point included in the three-dimensional points, same layer reference is performed when a total number of the one or more layers is one, the same layer reference including generating a prediction value of the attribute information of the current three-dimensional point by reference to attribute information of an other three-dimensional point included in a same layer as the current three-dimensional point.