H.264 Slice Priority Prediction Using Generalized Linear Model
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current video transmission over wireless networks is prone to degradation and distortion due to channel errors, which are exacerbated by the computational intensity of existing methods for predicting cumulative mean squared error (CMSE) in H.264 AVC video slices, leading to delays and inefficiencies in error propagation and quality management.
Innovation Solution
A low-complexity, low-delay generalized linear model (GLM) is developed to predict CMSE using video factors extracted during encoding, such as motion vectors and residual energy, allowing for real-time slice prioritization and error protection without requiring decoding of future frames, and is applied at both GOP and frame levels to optimize video quality.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If CMSE is computed for each slice loss to accurately measure distortion, then measurement precision is improved, but device complexity and loss of time increase due to requiring decoding of all frames in a GOP
Solution Approach 1:
The patent creates a simplified copy of the CMSE computation process by using a prediction model that replicates the essential distortion measurement functionality without requiring full GOP decoding. The model uses linear combination of video factors to predict CMSE values, providing accurate distortion measurement while avoiding the computational burden of actual decoding operations.
Solution Approach 2:
The patent performs preliminary action by pre-computing video factors during the encoding phase and storing them for later use. This allows the prediction model to quickly estimate CMSE values without needing to decode frames during transmission or reception, thereby reducing both complexity and delay while maintaining measurement precision.
2Measurement precision
If CMSE is computed for each slice loss to accurately measure distortion, then measurement precision is improved, but loss of time increases due to requiring decoding of all frames in a GOP
Solution Approach 1:
The prediction model creates a time-efficient copy of the CMSE measurement process by using pre-extracted video factors and linear combination calculations. This approach replicates the essential distortion assessment functionality instantaneously without the time-consuming GOP decoding process, enabling real-time quality monitoring and slice prioritization.
Solution Approach 2:
By extracting and storing video factors during the encoding phase (preliminary action), the system eliminates the need for time-consuming decoding operations during transmission and reception. The pre-computed factors are readily available for the prediction model to quickly estimate CMSE values, significantly reducing computational delay while maintaining measurement accuracy.
3Measurement precision
If video factors requiring decoding of future frames are used in the model, then prediction accuracy is improved, but device complexity and loss of time increase
Solution Approach 1:
The patent applies preliminary action by extracting and storing all necessary video factors during the current frame's encoding phase. This ensures that the prediction model has access to sufficient information from already-encoded frames without needing to wait for or decode future frames, thereby eliminating prediction delay while maintaining accurate CMSE estimation through the use of pre-computed motion vectors, residuals, and other video factors.
Data Source
AI summary
The invention relates to systems and methods for prioritizing video slices of H.264 video bitstream comprising: a memory storage and a processing unit coupled to the memory storage, wherein the processing unit operates to execute a low complexity scheme to predict the expected cumulative mean squared error (CMSE) contributed by the loss of a slice of H.264 video bitstream, wherein the processing unit operates to execute a series of actions comprising assigning each slice a predicted value according to the low complexity scheme; extracting video parameters during encoding process, said video parameters; and using a generalized linear model to model CMSE as a linear combination of the video parameters, wherein the video parameters are derived from analytical estimations by using a Generalized Linear Model (GLM) over a video database, encompassing videos of different characteristics such as high and low motion, camera panning, zooming and still videos, further comprising wherein the GLM is constructed in a training phase as follows: determining the distribution of the computed CMSE to be a Normal distribution with the Identity link function; sequentially adding covariates using the forward selection technique where by the best model is evaluated at each stage using the Akaike's Information Criterion (AIC); the training phase of the model generates regression coefficients; the final model is validated through the testing phase by predicting the CMSE for different video sequences, not in the training database; and by using the regression coefficients, the expected CMSE values are predicted for each slice.