Inter Prediction Encoding with Reconstructed Frame Fusion
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing AI video compression technologies suffer from accumulated errors in inter encoding processes, leading to deteriorated compression performance and time sequence instability due to errors introduced in prediction and entropy encoding of residual information.
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 encoding to compensate for motion complexity, thereby reducing errors and ensuring time sequence stability.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Loss of substance
If inter encoding is performed to reduce bitrate overheads, then bandwidth and storage requirements are reduced, but accumulated errors increase and compression performance deteriorates
Solution Approach 1:
The patent implements a feedback mechanism where the encoder sends reconstructed picture information and motion complexity parameters to the decoder, and the decoder uses this feedback to adjust its decoding strategy. This feedback loop allows the system to maintain compression performance stability by compensating for accumulated errors through adaptive bitrate allocation based on motion complexity assessment.
Solution Approach 2:
The patent dynamically changes encoding parameters based on motion complexity assessment. The encoder determines motion complexity parameters for different picture regions and adjusts bitrate allocation accordingly, using higher bitrates for high-motion regions and lower bitrates for low-motion regions. This parameter adaptation prevents error accumulation in critical regions while maintaining overall compression efficiency.
2Productivity
If AI-based inter encoding is used to improve compression effect, then bandwidth and storage are reduced, but errors in prediction and entropy encoding accumulate over GOP
Solution Approach 1:
The patent applies local quality assessment by dividing the picture into different regions and evaluating motion complexity for each region independently. Different encoding precision and bitrate allocation are applied to different regions based on their motion characteristics. High-motion regions receive higher encoding precision to maintain prediction accuracy, while low-motion regions use lower precision to save bandwidth, thus preventing error accumulation in critical areas.
Solution Approach 2:
The patent introduces dynamic adaptation where encoding parameters are not fixed but change based on the actual motion complexity of each picture and region. The system dynamically adjusts bitrate allocation, prediction mode selection, and entropy encoding parameters according to real-time motion assessment, allowing the compression system to maintain high prediction accuracy in challenging regions while optimizing overall efficiency.
3Loss of substance
If conventional video compression is used to control bandwidth, then transmission delay and storage costs are managed, but video quality requirements for ultra HD are not met
Solution Approach 1:
The patent employs parameter changes by assessing motion complexity and dynamically adjusting encoding parameters including bitrate allocation, transformation block sizes, and quantization parameters. This allows the system to achieve ultra HD quality where needed while maintaining bandwidth efficiency overall, as high-quality encoding is applied selectively rather than uniformly across all picture content.
Solution Approach 2:
The patent implements local quality enhancement by identifying high-motion regions that require higher encoding quality to maintain video fidelity. These regions receive enhanced encoding resources and higher bitrate allocation, while low-motion regions use more aggressive compression. This localized quality approach meets ultra HD requirements for important content while controlling overall bandwidth consumption.
Data Source
Figure 1A
Figure 1B
Figure 1C
AI summary
Embodiments of this application provide an encoding method, a decoding method, and an electronic device. The method includes: performing inter prediction on a coding unit in a current frame, to obtain first prediction information of the coding unit; determining residual information of the coding unit based on the first prediction information and an original picture of the coding unit; determining a preset parameter of the coding unit, where the preset parameter is used to fuse reconstructed picture information of a reference unit and reconstructed picture information of the coding unit, to obtain a target reconstructed picture of the coding unit, the reconstructed picture information of the reference unit includes a target reconstructed picture of the reference unit or feature information of the target reconstructed picture of the reference unit, and the reconstructed picture information of the coding unit includes an initial reconstructed picture of the coding unit or feature information of the initial reconstructed picture of the coding unit; and encoding the residual information, and encoding the preset parameter. In this way, an accumulated error can be reduced, and time sequence stability is ensured.