POC Derivation in Image Decoding for Stable Reference Prediction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The increasing demand for high-resolution and high-quality images/videos, particularly in immersive media formats like VR and AR, leads to higher transmission and storage costs due to increased bit amounts, necessitating a more efficient image/video compression technique.
Innovation Solution
The method and apparatus enhance image coding efficiency by improving POC decoding efficiency, restricting the POC value between current and reference pictures, and avoiding the use of deleted reference pictures, using POC information and non-reference picture flags to optimize inter prediction.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Manufacturing precision
If high-resolution and high-quality images/videos are transmitted or stored, then image quality is improved, but transmission cost and storage cost increase
Solution Approach 1:
The patent extracts only the essential POC information (maximum LSB value) and reference picture flags from the bitstream, deriving other POC values through calculation rather than direct transmission. This removes redundant data while preserving the ability to perform accurate inter prediction, thereby reducing bit amount while maintaining image quality.
Solution Approach 2:
The patent performs preliminary derivation of POC values during the decoding process by calculating them from the transmitted maximum LSB value and reference picture POC values. This preliminary computation eliminates the need to transmit complete POC information for all reference pictures, reducing transmission data while ensuring accurate prediction.
2Productivity
If POC information is decoded using conventional methods, then decoding is performed, but error occurrence increases when reference pictures are deleted
Solution Approach 1:
The patent introduces a non-reference picture flag that is set in advance during encoding to mark pictures that will be deleted or removed from the reference picture list. During decoding, this flag allows the system to proactively identify and exclude invalid reference pictures before they cause prediction errors, preventing reliability issues before they occur.
Solution Approach 2:
The patent implements a feedback mechanism where the non-reference picture flag information is continuously monitored during the decoding and reference picture list construction process. When a picture marked with the non-reference flag is encountered, the system adjusts the reference picture list accordingly, providing real-time correction to maintain decoding reliability.
3Measurement precision
If complete POC information is transmitted for all reference pictures, then POC accuracy is improved, but transmission data amount increases
Solution Approach 1:
The patent extracts only the critical maximum LSB value of POC and reference picture flags for transmission. The complete POC accuracy is maintained by deriving individual reference picture POC values during decoding using the formula: POC value of current picture minus (POC value of previous picture minus POC value of reference picture), eliminating the need to transmit redundant POC data while preserving precision.
Solution Approach 2:
The patent changes the representation parameter from transmitting complete POC values for all reference pictures to transmitting only the maximum LSB value and differential POC information. This parameter transformation maintains the ability to reconstruct accurate POC values while significantly reducing the transmission data amount.
Data Source
AI summary
An image decoding method according to the present document comprises a step of deriving POC values for the current picture and reference pictures on the basis of POC information, wherein the POC information includes a maximum LSB value of POC, information about the reference pictures includes a non-reference picture flag related to whether a picture is not used as a reference picture, the value of the non-reference picture flag of the previous picture used for deriving the POC value of the current picture is zero, and the difference between the POC values of the current picture and the previous picture is less than half of the maximum LSB value of POC.