Skip Encoding With Filtered Intra Prediction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing image coding methods suffer from blocking artifacts, quantization errors, and inefficient intra prediction due to the lack of deblocking filtering and spatially limited prediction, which affects the efficiency of coding and decoding processes.
Innovation Solution
Implementing skip coding that generates prediction signals similar to coding object signals using filtered reconstructed signals and skips residual signals, employing various filters like low pass, deblocking, adaptive loop, interpolation, and noise removing filters to enhance intra prediction.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If deblocking filtering is not applied to spatially adjacent reconstructed signals during intra prediction, then device complexity is reduced, but blocking artifacts occur degrading image quality
Solution Approach 1:
The patent applies deblocking filtering to spatially adjacent reconstructed signals before they are used for intra prediction. This preliminary filtering action prevents blocking artifacts from occurring in the first place, rather than attempting to remove them later. By pre-filtering the reference signals, the invention eliminates the harmful blocking effects while maintaining device complexity at acceptable levels.
2Device complexity
If only pixels in integer pel unit are used for adjacent reconstructed signals, then device complexity is reduced, but measurement precision of prediction deteriorates
Solution Approach 1:
The patent transitions from using only integer pel unit pixels to utilizing sub-pel precision values for adjacent reconstructed signals. This dimensional change in precision allows the prediction process to access finer granularity pixel data, thereby improving measurement precision and prediction accuracy without excessively increasing device complexity.
3Device complexity
If only object signals of coding and adjacent reconstructed signals are used for intra prediction, then device complexity is reduced, but adaptability of prediction deteriorates
Solution Approach 1:
The patent enhances the intra prediction process by incorporating filtered reconstructed signals from multiple spatial locations and temporal references. This multi-functionality approach allows the prediction system to adapt to various coding scenarios and image characteristics, improving versatility and adaptability while managing device complexity through efficient signal selection and filtering.
4Manufacturing precision
If residual signals are coded to capture differences between original and prediction signals, then manufacturing precision of reconstruction is improved, but productivity of coding process deteriorates
Solution Approach 1:
The patent implements skip coding mode that selectively skips the encoding of residual signals when the prediction accuracy is sufficient. By detecting cases where prediction closely matches the original signal, the system rushes through these regions without performing full residual coding, thereby maintaining reconstruction precision for important details while significantly improving coding productivity and speed for redundant regions.
Data Source
AI summary
The present invention relates to an apparatus and method for encoding and decoding an image by skip encoding. The image-encoding method by skip encoding, which performs intra-prediction, comprises: performing a filtering operation on the signal which is reconstructed prior to an encoding object signal in an encoding object image; using the filtered reconstructed signal to generate a prediction signal for the encoding object signal; setting the generated prediction signal as a reconstruction signal for the encoding object signal; and not encoding the residual signal which can be generated on the basis of the difference between the encoding object signal and the prediction signal, thereby performing skip encoding on the encoding object signal.