Collocated Picture Selection for Temporal Prediction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
In High Efficiency Video Coding (HEVC), the selection of an optimal reference picture for temporal prediction is often inefficient due to the reliance on predefined lists (list0 and list1), which may not always contain the closest temporal reference picture, leading to suboptimal motion vector prediction and increased coding overhead.
Innovation Solution
A method that determines the closest temporal reference picture to the current picture by analyzing the temporal distance between the current picture and reference pictures in both list0 and list1, allowing for implicit or explicit selection of the collocated picture for use in temporal prediction processes, thereby improving coding efficiency.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of operation
If the first reference picture in list0 or list1 is selected as the collocated picture, then the selection process is simple, but the temporal distance may not be minimal leading to suboptimal motion vector prediction
Solution Approach 1:
The patent changes the selection criterion from a fixed rule (first reference picture) to a dynamic parameter-based selection (reference picture with minimum temporal distance). The temporal distance parameter is calculated and compared across multiple reference pictures to identify the optimal collocated picture, thereby resolving the contradiction between simple selection and accurate temporal reference.
2Productivity
If motion vectors from distant temporal reference pictures are used, then the reference picture selection is easier, but the difference from current motion vector increases reducing coding efficiency
Solution Approach 1:
The patent replaces the conventional fixed-rule selection mechanism with a calculation-based temporal distance analysis system. By computing temporal distances and selecting the minimum, the system substitutes a simple but inaccurate mechanical selection process with a more complex but precise analytical approach, improving motion vector prediction accuracy and coding efficiency.
Data Source
AI summary
In one embodiment, a method determines a current picture and determines at least one of a first list of reference pictures and a second list of reference pictures for the current picture. A temporal distance is analyzed between the current picture and at least a portion of the reference pictures in the at least one of a first list of reference pictures and a second list of reference pictures. The method then selects reference picture that is closest in temporal distance from the current picture in the analyzed at least a portion of the reference pictures. The selected reference picture is set as a collocated picture for the current picture and the selected reference picture is used in a temporal prediction process for the current picture.


