Adaptive Search Range Video Encoding Reduces Power
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Video encoding devices using fixed search ranges for inter-prediction operations consume excessive power and time due to the lengthy process of obtaining optimum motion vectors, especially when encoding videos according to standards like MPEG-4 and H.264.
Innovation Solution
A video encoding device and method that adaptively determines a search range for inter-prediction operations by using accumulated motion vectors derived from optimum motion vectors obtained in the base layer, allowing for optimized search ranges in hierarchical layers, thereby reducing unnecessary search operations and power consumption.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If a fixed large search range is used for inter-prediction operation, then the optimum motion vector can be obtained accurately, but the encoding time and power consumption increase significantly
Solution Approach 1:
The patent applies dynamics by making the search range adaptive rather than fixed. The search range is dynamically adjusted based on the hierarchical layer being encoded and the accumulated motion vector information from previously decoded pictures. This allows the search range to be large when needed for accuracy and small when sufficient motion information is already available, resolving the contradiction between motion vector accuracy and encoding time
Solution Approach 2:
The patent changes the parameter of search range size based on encoding conditions. By calculating an accumulated motion vector from decoded pictures and using it to determine an adaptive search range for hierarchical layers, the system optimizes the balance between obtaining accurate motion vectors and minimizing encoding time. The search range parameter is modified according to the specific encoding context rather than remaining constant
2Measurement precision
If a fixed large search range is used for inter-prediction operation, then the optimum motion vector can be obtained accurately, but the power consumption increases
Solution Approach 1:
The system dynamically adjusts the search range based on the hierarchical layer and accumulated motion vector information. This dynamic adaptation ensures that the search range is only as large as necessary to find accurate motion vectors, reducing unnecessary computational operations and thereby lowering power consumption while maintaining motion vector accuracy when required
Solution Approach 2:
The search range parameter is changed adaptively based on encoding conditions in hierarchical layers. By utilizing accumulated motion vector information from base layer pictures, the system modifies the search range parameter to minimize computational workload and power consumption while ensuring adequate motion vector precision is achieved
3Adaptability or versatility
If inter-prediction operation is performed on hierarchical layer pictures with reference to base layer pictures, then motion information can be effectively utilized, but the search complexity increases
Solution Approach 1:
The patent applies preliminary action by calculating the accumulated motion vector from base layer pictures before performing inter-prediction on hierarchical layer pictures. This pre-computed motion information is then used to define the adaptive search range, simplifying the subsequent search process and reducing complexity while effectively utilizing motion information across layers
Solution Approach 2:
The accumulated motion vector acts as an intermediary between base layer pictures and hierarchical layer pictures. It bridges the two layers by providing pre-processed motion information that guides the search process, reducing the complexity of direct inter-layer prediction while maintaining effective motion information utilization
Data Source
AI summary
A device and method for encoding a picture sequence including a group of pictures (GOP) having a base layer and a hierarchical layer. The device includes a first unit and a second unit. The first unit is configured to perform an inter-prediction operation on first unit blocks of a first picture in the base layer with reference to first reference picture in the base layer using a first search range, and to obtain first optimum motion vectors corresponding to the first unit blocks of the first picture. The second unit is configured to determine a second search range using the first optimum motion vectors. The first unit is configured to perform the inter-prediction operation on second unit blocks of a second picture in the hierarchical layer using the second search range.


