Scalable Video Reference Picture List Construction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current spatial enhancement layer reference picture list construction in scalable video coding is complex and requires different reordering commands for base and enhancement layers, especially when the enhancement layer has a higher frame rate, leading to increased data transmission and processing complexity.
Innovation Solution
A new decoding process that uses a single syntax flag to simplify the construction of reference picture lists by predicting enhancement layer lists from base layer lists, reducing the need for complicated reordering commands and allowing for error concealment in case of data loss.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If different RPLR commands are used for base layer and enhancement layer, then the reference picture lists can be accurately constructed for each layer, but the data transmission and processing complexity increases
Solution Approach 1:
The patent merges the reference picture list construction processes of the base layer and enhancement layer by introducing a flag that enables the enhancement layer to inherit and reuse the base layer's reference picture lists. This combining approach eliminates the need for separate RPLR commands in the enhancement layer, reducing processing complexity while maintaining list accuracy through the inheritance mechanism.
Solution Approach 2:
The patent makes the base layer's reference picture lists universally applicable to both the base layer and enhancement layer. By designing the system so that enhancement layer pictures can use the same reference lists as the base layer (when the flag is set), the solution achieves multi-functionality where a single set of reference lists serves dual purposes, reducing the need for layer-specific processing.
2Measurement precision
If RPLR commands are transmitted for enhancement layer, then reference picture lists can be reordered accurately, but data transmission volume increases
Solution Approach 1:
The patent extracts the RPLR command transmission requirement from the enhancement layer by introducing a flag that, when set, eliminates the need for separate RPLR commands in the enhancement layer. The enhancement layer inherits the base layer's reference picture lists directly, taking out the redundant command transmission while preserving the necessary reordering information through the inheritance mechanism.
Solution Approach 2:
The patent uses copying by having the enhancement layer copy the base layer's reference picture lists when the flag is set. Instead of transmitting separate RPLR commands for the enhancement layer, the system copies the existing base layer lists, reducing data transmission volume while maintaining the necessary reference picture information for accurate decoding.
3Quantity of substance
If enhancement layer uses same RPLR commands as base layer, then data transmission is reduced, but reference picture list accuracy may be compromised when frame rates differ
Solution Approach 1:
The patent introduces dynamics by making the reference picture list construction adaptive based on the frame rate relationship between layers. When frame rates are equal, the enhancement layer uses the same RPLR commands as the base layer. When frame rates differ, the system dynamically adjusts by inheriting the base layer's reference lists through the flag mechanism, ensuring accuracy is maintained while data transmission is minimized.
Solution Approach 2:
The patent changes the parameter of reference picture list inheritance based on the frame rate ratio between base and enhancement layers. When the frame rates are equal, traditional separate RPLR commands are used. When the enhancement layer frame rate is higher (e.g., twice the base layer), the system changes to using the inheritance flag, allowing the enhancement layer to use the base layer's reference lists with adjusted picture order counts, thereby maintaining accuracy while reducing data transmission.
Data Source
AI summary
In video coding, prediction of P- and B-frames is based on reference frames, which are indicated by reference picture lists. Scalable video coding (SVC) requires this for both, base-layer and enhancement-layer. The invention describes a decoding process for reference picture list construction for spatial enhancement layer. With just one flag of syntax modification, it provides simpler and direct reference lists construction process. By using this, complicated RPLR and other syntaxes can be skipped, and the RPLR process for spatial enhancement layer can also be avoided. The process also can be used as for error concealment when the spatial enhancement layer slice is lost.