HEVC Reference Frame Selection via Spatial Correlation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The existing methods for selecting an optimal reference frame in HEVC coding require individual testing of candidate reference frames, leading to high computation costs and long processing times, which negatively impact coding efficiency.
Innovation Solution
The method involves determining the optimal reference frame for prediction units by leveraging the correlation between spatially adjacent prediction units in the same or different layer depths, reducing the need for ergodic rate distortion cost calculations, especially for units in the minimum layer depth.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If ergodic selection method is used to test all candidate reference frames individually, then optimal reference frame selection accuracy is improved, but computation cost and processing time increase significantly
Solution Approach 1:
The patent performs rate distortion cost calculation for candidate reference frames in advance, before the actual inter-frame prediction process. By pre-calculating and storing the cost values, the system avoids repeated calculations during encoding, thus maintaining selection accuracy while reducing processing time.
Solution Approach 2:
The patent combines the reference frame selection process with the rate distortion cost calculation process. Instead of separately testing each candidate frame and then selecting, the system integrates both operations by calculating costs for all candidates and directly selecting the minimum, reducing overall computation overhead.
2Measurement precision
If ergodic selection method is used to test all candidate reference frames individually, then reference frame selection accuracy is improved, but coding efficiency deteriorates
Solution Approach 1:
The patent calculates rate distortion costs for all candidate reference frames in advance, before the encoding process begins. This preliminary calculation allows the system to quickly select the optimal reference frame during encoding without performing repeated computations, thereby maintaining accuracy while improving coding efficiency.
Solution Approach 2:
The patent extracts the rate distortion cost calculation from the main encoding loop and performs it separately in advance. By separating this computationally intensive operation from the time-critical encoding process, the system maintains selection accuracy while preventing it from becoming a bottleneck to coding efficiency.
3Manufacturing precision
If multiple CUs and PUs are processed with full ergodic search, then prediction accuracy is improved, but computation cost increases
Solution Approach 1:
The patent performs rate distortion cost calculation for all candidate reference frames in advance, before processing multiple CUs and PUs. By pre-computing these costs and storing them, the system avoids repeated calculations when processing multiple coding units and prediction units, thus maintaining prediction accuracy while reducing overall computation cost.
Solution Approach 2:
The patent calculates rate distortion costs for all candidate reference frames in advance, which is more than the minimum required for each individual PU. This excessive preliminary calculation benefits multiple PUs and CUs that share the same candidate frames, reducing total computation cost across the entire encoding process.
Data Source
AI summary
A method, an apparatus and a coder for selecting an optimal reference frame in HEVC coding are disclosed. The method includes obtaining optimal reference frames of prediction units spatially adjacent to a current prediction unit, and determining an optimal reference frame of the current prediction unit based on a correlation between optimal reference frames of the spatially adjacent prediction units in a same layer depth for prediction unit(s) of a minimum layer depth. In this way, an ergodic rate distortion cost calculation is not needed to be performed at least for the prediction unit(s) of the minimum layer depth, and the corresponding optimal reference frame is directly determined using the correlation, thus reducing the amount of calculations and enhancing the coding efficiency.