Image Coding With Bidirectional Motion Vector Candidate Lists
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional image coding methods limit the use of motion vectors from only one prediction direction, leading to suboptimal motion vector derivation and reduced coding efficiency.
Innovation Solution
An image coding method that adds adjacent motion vectors from both the first and second reference picture lists as candidates for predicting motion vectors, increasing the options for deriving a suitable predicted motion vector.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If motion vectors are derived using only one prediction direction, then the complexity of motion vector derivation is reduced, but the coding efficiency deteriorates
Solution Approach 1:
The motion vector prediction process is segmented into multiple prediction directions (first prediction direction and second prediction direction), each with its own candidate list. This allows the system to evaluate multiple directional candidates separately and select the best one, improving coding efficiency without overwhelming complexity.
Solution Approach 2:
The invention extends the motion vector prediction from a single direction to multiple dimensions by introducing both first and second prediction directions. Each direction has its own reference picture list and candidate motion vectors, creating a multi-dimensional search space that improves prediction accuracy and coding efficiency.
2Measurement precision
If motion vectors from both first and second reference picture lists are used as candidates, then the accuracy of predicted motion vector is improved, but the complexity of motion vector derivation increases
Solution Approach 1:
The candidate motion vectors are segmented into two distinct groups: those from the first reference picture list and those from the second reference picture list. Each group is evaluated independently in its respective prediction direction, allowing accurate selection without overwhelming complexity.
Solution Approach 2:
The system dynamically selects which reference picture list to use based on the prediction direction and current block characteristics. The apparatus can adaptively choose between first and second prediction directions, making the motion vector derivation process flexible and efficient.
Data Source
AI summary
An image coding method includes: adding, to a candidate list, a first adjacent motion vector as a candidate for a predicted motion vector to be used for coding the current motion vector; selecting the predicted motion vector from the candidate list; and coding the current motion vector, wherein in the adding, the first adjacent motion vector indicating a position in a first reference picture included in a first reference picture list is added to the candidate list for the current motion vector indicating a position in a second reference picture included in a second reference picture list.