Deinterlacing Algorithm Reducing GPU Power and Silicon Size

Solution Overview

Problem

Conventional motion adaptive edge-directional deinterlacing methods require dedicated hardware, consume excessive power, and are not suitable for low-end graphical processing units (GPUs), leading to inefficiencies in deinterlacing high-quality video in real time.

Innovation Solution

A method and apparatus that construct a temporary frame using pixels from a current field, placeholder pixels in motion areas, and pixels from an opposite field in still areas, with edge directional interpolation replacing placeholder pixels to generate a reconstructed frame, leveraging existing hardware more efficiently.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Manufacturing precision

If dedicated hardware is used for motion adaptive edge-directional deinterlacing, then deinterlacing quality is improved, but device complexity and power consumption increase

Engineering Contradiction:

Improvedeinterlacing qualityVSAvoidhardware complexity

Core Design Contradiction:

Manufacturing precisionVSDevice complexity

Solution Approach 1:

The patent uses software-based deinterlacing algorithms that copy and process pixel data through multiple passes, replicating the functionality of dedicated hardware through programmable operations on standard GPUs

Inventive Principle:

Principle #26Copying

Solution Approach 2:

The patent replaces dedicated hardware circuits with software algorithms executed on general-purpose GPUs, substituting mechanical/electrical systems with computational software solutions that achieve the same deinterlacing function

Inventive Principle:

Principle #28Mechanics substitution (Replace mechanical system)

2Manufacturing precision

If dedicated hardware is used for motion adaptive edge-directional deinterlacing, then deinterlacing quality is improved, but power consumption increases

Engineering Contradiction:

Improvedeinterlacing qualityVSAvoidpower consumption

Core Design Contradiction:

Manufacturing precisionVSUse of energy by moving object

Solution Approach 1:

The patent uses software-based deinterlacing algorithms that copy and process pixel data through multiple passes, replicating the functionality of dedicated hardware through programmable operations on standard GPUs

Inventive Principle:

Principle #26Copying

Solution Approach 2:

The patent replaces dedicated hardware circuits with software algorithms executed on general-purpose GPUs, substituting mechanical/electrical systems with computational software solutions that achieve the same deinterlacing function

Inventive Principle:

Principle #28Mechanics substitution (Replace mechanical system)

3Productivity

If placeholder pixels are used in motion areas, then processing efficiency is improved, but interpolation accuracy may be reduced

Engineering Contradiction:

Improveprocessing efficiencyVSAvoidinterpolation accuracy

Core Design Contradiction:

ProductivityVSManufacturing precision

Solution Approach 1:

The patent segments the deinterlacing process into multiple passes: first identifying motion areas and placing placeholder pixels for efficiency, then performing edge-directional interpolation in a second pass to restore accuracy where needed

Inventive Principle:

Principle #1Segmentation

Solution Approach 2:

The patent performs preliminary motion detection and placeholder pixel placement before the final interpolation step, preparing the data structure in advance to guide subsequent accurate interpolation only where necessary

Inventive Principle:

Principle #10Preliminary action

Data Source

PatentUS8259228B2Method and apparatus for high quality video motion adaptive edge-directional deinterlacing

Publication Date: 2012.09.04 ATI TECHNOLOGIES ULC

AI summary

A method for deinterlacing video includes constructing a temporary frame of deinterlaced video based on a first (i.e., current) field of interlaced video, wherein the temporary frame includes pixels in lines of the temporary frame associated with the first field of interlaced video, placeholder pixels in identified areas of motion in lines of the frame associated with a missing field of interlaced video, and pixels from an opposite field of polarity of interlaced video in areas without motion. The method further includes replacing the placeholder pixels in the identified areas of motion with pixels interpolated using an edge direction interpolation scheme based on pixels in the first field of interlaced video, resulting in a reconstructed frame. In one example, a motion adaptive interpolator may construct the temporary frame, and an edge directional interpolator may generate the reconstructed/deinterlaced the frame.