Video Decoding Merge Candidate List Configuration

The merge candidate list is configured dynamically based on the number of available candidates. When the number of merge candidates is greater than 1 and less than a predetermined maximum, an additional merge candidate is generated by averaging the first and second merge candidates. This dynamic configuration allows the system to adapt the list size and content to the specific conditions of each current block, thereby improving prediction accuracy without requiring a fixed complex configuration for all cases.

The system changes the parameter of merge candidate list size and content based on the number of available candidates. By conditionally adding an additional merge candidate (when 1 < number of candidates < maximum), the system optimizes the prediction process for different scenarios, resolving the contradiction between maintaining simple fixed configuration and achieving high prediction accuracy across varying conditions.

The generation of additional merge candidates is performed dynamically only when needed (when the number of candidates is between 1 and the maximum). This conditional approach ensures that computational resources are spent on generating additional candidates only in scenarios where they will improve prediction accuracy, rather than always performing the computation, thus maintaining encoding/decoding speed while improving accuracy when beneficial.

The system performs the averaging operation to generate additional merge candidates using the existing first and second merge candidates already in the list. This self-service approach reuses available data without requiring external complex computations or additional hardware resources, minimizing the computational overhead while still improving prediction accuracy.