Edge Anti-Aliasing via Border Region and Linear Filtering

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing anti-aliasing techniques in digital signal processing require significant system resources and can reduce overall system performance, leading to noticeable defects in displayed images due to aliasing and distortion.

Innovation Solution

The technique involves modifying graphical objects by adding a border region around them, selecting its color based on the background, and using linear filtering, which allows for efficient edge anti-aliasing by smoothing transitions between the object and background, potentially eliminating the need for alpha blending when the background is solid.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Manufacturing precision

If traditional anti-aliasing signal processing techniques are used, then image quality is improved, but system resources are consumed and overall system performance is reduced

Engineering Contradiction:
Improveimage qualityVSAvoidsystem performance
Core Design Contradiction:
Manufacturing precisionVSProductivity

Solution Approach 1:

The patent extracts the anti-aliasing function from traditional signal processing techniques and implements it through geometric modification of graphical objects by adding border regions. This separates the anti-aliasing function from resource-intensive filtering operations, achieving image quality improvement without significant system resource consumption.

Inventive Principle:
Principle #2Taking out (Extraction)

Solution Approach 2:

The patent segments the graphical object into the original object and an added border region. This segmentation allows the border to handle anti-aliasing independently through linear filtering, reducing the computational burden on the entire image processing system while maintaining image quality.

Inventive Principle:
Principle #1Segmentation

2Manufacturing precision

If traditional anti-aliasing signal processing is applied, then aliasing reduction is achieved, but computational resources and memory bandwidth are consumed

Engineering Contradiction:
Improvealiasing reductionVSAvoidcomputational resources
Core Design Contradiction:
Manufacturing precisionVSUse of energy by moving object

Solution Approach 1:

The patent performs preliminary action by adding border regions to graphical objects before rendering. This pre-processing step prepares the object for efficient linear filtering, reducing the need for complex real-time anti-aliasing computations and thereby lowering computational resource usage during image display.

Inventive Principle:
Principle #10Preliminary action

3Manufacturing precision

If linear filtering is used on modified graphical objects, then edge smoothness is improved and stairstepping is reduced, but additional processing steps are required

Engineering Contradiction:
Improveedge smoothnessVSAvoidprocessing steps
Core Design Contradiction:
Manufacturing precisionVSDevice complexity

Solution Approach 1:

The patent merges the anti-aliasing function with the existing linear filtering capability of the graphics hardware. By combining border region addition with standard linear filtering operations, the system achieves edge smoothness improvement without requiring separate dedicated anti-aliasing processing steps.

Inventive Principle:
Principle #5Merging (Combining)

Data Source

PatentUS7849414B2Edge anti-aliasing
Publication Date: 2010.12.07 APPLE INC
  • US7849414B2 patent drawing
  • US7849414B2 patent drawing
  • US7849414B2 patent drawing

AI summary

Techniques for anti-aliasing of graphical objects. A graphical object to be displayed on a display image may be modified to include a border region at least partially surrounding the object. The color of the border region may be selected based on the color of the background over which the graphical object is to be displayed. The graphical object is linearly filtered. The linearly filtered graphical object is displayed on a display device.