Image Processing Apparatus Offset Control for Reduced Circuit Complexity
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The processing amount of the SAO filter tends to increase when an optimum mode and offset value are set, leading to increased circuit size or power consumption.
Innovation Solution
An image processing apparatus and method that includes a filter processing section for applying an offset to decoded image pixels and a control unit to control the filter process based on encoding parameters, reducing the processing amount by optimizing the filter process.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Manufacturing precision
If an optimum mode and offset value are set for the SAO filter, then coding efficiency is improved, but processing amount increases leading to larger circuit size or higher power consumption
Solution Approach 1:
The patent changes the parameter range for offset values based on the quantization parameter (QP). When QP is large (coarse quantization), the patent limits offset values to a smaller range (e.g., -3 to 3) compared to when QP is small. This parameter adaptation reduces the number of cost calculation candidates without significantly impacting coding efficiency in high-QP scenarios, thereby reducing processing amount and circuit complexity.
2Manufacturing precision
If an optimum mode and offset value are set for the SAO filter, then coding efficiency is improved, but power consumption increases
Solution Approach 1:
The patent dynamically adjusts the offset value range parameter based on the quantization parameter (QP). By limiting the offset search range to smaller values when QP is large, the number of cost calculations required is reduced, which directly reduces the processing amount and consequently lowers power consumption in the filter processing section while maintaining adequate coding efficiency.
3Measurement precision
If cost calculation is performed for all possible offset values, then filtering accuracy is improved, but processing amount increases
Solution Approach 1:
The patent adapts the offset value range parameter based on the quantization parameter (QP). When QP is large, the patent uses a smaller offset range (e.g., -3 to 3), reducing the number of cost calculations from potentially 32 candidates to just 7 candidates. This parameter-based optimization maintains filtering accuracy where it matters most while significantly reducing processing amount in high-QP scenarios.
Solution Approach 2:
The patent applies partial action by performing cost calculation for a limited set of offset values rather than all possible values. By selecting a subset of offset candidates based on QP, the patent achieves adequate filtering performance without the excessive processing required to evaluate every possible offset value, particularly when quantization is already coarse.
Data Source
AI summary
There is provided an image processing apparatus and an image processing method which are capable of reducing a processing amount of a cost calculation, the image processing apparatus including: a filter processing section configured to perform a filter process of applying an offset to a pixel of a decoded image that is decoded; and a control unit configured to control a mode used when the filter process is performed and an offset for the mode according to an encoding parameter used when an image is encoded.