Inter Prediction Fusion for Stable AI Video Encoding
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing AI video compression technologies suffer from accumulated errors and deteriorating compression performance due to inter encoding processes, particularly in high-definition and ultra-high-definition video encoding, leading to instability in time sequence and increased bitrate overheads.
Innovation Solution
An encoding method that fuses reconstructed picture information of a coding unit with a reference unit based on a preset parameter, adjusting bitrate and optical flow encoding to compensate for motion complexity, thereby reducing errors and improving compression performance.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Quantity of substance
If inter encoding is performed to reduce bitrate overheads, then bitrate efficiency is improved, but accumulated error increases and compression performance deteriorates
Solution Approach 1:
The patent applies preliminary action by performing error compensation before the accumulated error significantly deteriorates compression performance. The error compensation mechanism is integrated into the inter encoding process, proactively correcting prediction errors as they accumulate across GOP groups, thereby maintaining compression performance stability while preserving bitrate efficiency.
Solution Approach 2:
The patent implements feedback through an error compensation mechanism that monitors and corrects accumulated errors in the inter encoding process. By feeding back compensation information based on detected prediction errors, the system maintains compression performance stability without sacrificing bitrate efficiency, resolving the contradiction between bitrate overhead reduction and performance stability.
2Productivity
If AI-based inter encoding is used to improve compression effect, then compression ratio is improved, but error accumulation increases over time
Solution Approach 1:
The patent employs feedback through an error compensation mechanism that tracks prediction errors accumulated during AI-based inter encoding. By continuously monitoring and compensating for these errors across GOP groups, the system maintains high compression ratios while preventing information loss from error accumulation, thus resolving the contradiction between compression efficiency and error control.
Solution Approach 2:
The patent introduces an error compensation mechanism as an intermediary between the AI prediction process and the final encoding output. This intermediary component corrects prediction errors without significantly increasing bitrate, allowing the system to maintain high compression ratios while reducing information loss from error accumulation.
3Quantity of substance
If conventional video compression is used to control bandwidth, then bandwidth usage is reduced, but video quality requirements cannot be met
Solution Approach 1:
The patent applies parameter changes by transitioning from conventional video compression to AI-based inter encoding with error compensation. This parameter change in the encoding approach enables the system to maintain low bandwidth usage while significantly improving video quality through more accurate prediction and error correction mechanisms, thus resolving the contradiction between bandwidth efficiency and quality requirements.
Data Source
AI summary
Embodiments of this application provide an encoding method, a decoding method, and an electronic device. The method includes: determining a preset parameter of a coding unit in a current frame; performing inter prediction on the coding unit, to obtain first prediction information of the coding unit; fusing the first prediction information and reconstructed picture information of a reference unit based on the preset parameter, to obtain second prediction information of the coding unit, wherein the reconstructed picture information of the reference unit comprises a target reconstructed picture of the reference unit or feature information of the target reconstructed picture of the reference unit, and the reference unit is a unit in a reference frame corresponding to the current frame; determining residual information of the coding unit based on the second prediction information and an original picture of the coding unit; and encoding the residual information and the preset parameter.