Non-Linear Color Space Conversion for GPU Gradient Filtering

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing graphics processing systems face challenges in efficiently converting non-linear color spaces, particularly in converting non-linear YUV samples to linear space, which is computationally intense and not directly supported by hardware, and require complex filtering processes.

Innovation Solution

The proposed solution involves non-linear YUV conversion processes that filter in a non-linear space and convert to a linear space during the conversion process, utilizing up-sampling and color space conversion operations with non-linear values.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Ease of manufacture

If linear interpolation methods are used for color space conversion, then computational simplicity is maintained, but filtering artifacts and color banding occur in gradient regions

Engineering Contradiction:
Improvecomputational simplicityVSAvoidcolor accuracy
Core Design Contradiction:
Ease of manufactureVSManufacturing precision

Solution Approach 1:

The color space conversion is segmented into multiple passes: a first pass performs linear interpolation for most pixels, while a second pass applies non-linear filtering only to specific pixels identified as needing it (those in gradient regions). This segmentation allows the system to maintain computational simplicity for the majority of pixels while improving color accuracy for problematic regions.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

Instead of applying non-linear filtering to all pixels (excessive action), the patent applies it only partially to pixels that benefit most from it. The system identifies pixels in gradient regions and applies non-linear filtering selectively, avoiding the computational overhead of processing every pixel with the more complex algorithm.

Inventive Principle:
Principle #16Partial or excessive action

2Manufacturing precision

If non-linear filtering is applied to all pixels, then filtering artifacts are reduced, but processing time and computational resources increase significantly

Engineering Contradiction:
Improvecolor accuracyVSAvoidprocessing time
Core Design Contradiction:
Manufacturing precisionVSLoss of time

Solution Approach 1:

The pixel population is segmented into two groups: pixels processed with simple linear interpolation and pixels processed with non-linear filtering. The patent uses a hybrid approach where most pixels go through the fast linear path, while only pixels identified as being in gradient regions undergo the slower non-linear filtering process.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

Non-linear filtering is applied partially rather than universally. The system performs non-linear filtering on a subset of pixels that are most likely to exhibit filtering artifacts, thereby reducing the overall processing time while still achieving the benefit of artifact reduction where needed.

Inventive Principle:
Principle #16Partial or excessive action

3Productivity

If color space conversion is performed without filtering, then processing speed is maintained, but color banding artifacts appear in smooth gradient regions

Engineering Contradiction:
Improveprocessing speedVSAvoidcolor smoothness
Core Design Contradiction:
ProductivityVSManufacturing precision

Solution Approach 1:

The conversion process is segmented into two stages: a first conversion pass that processes all pixels quickly using linear interpolation, and a second conversion pass that applies non-linear filtering only to specific pixels. This segmentation maintains high processing speed for the majority of pixels while improving color smoothness for gradient regions.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The patent performs a preliminary color space conversion using linear interpolation for all pixels before applying non-linear filtering to specific pixels. This preliminary action establishes baseline color values that can then be refined selectively, maintaining processing efficiency while improving color smoothness where needed.

Inventive Principle:
Principle #10Preliminary action

4Manufacturing precision

If high precision color conversion is applied to all pixels, then color accuracy is improved, but power consumption and computational overhead increase

Engineering Contradiction:
Improvecolor accuracyVSAvoidpower consumption
Core Design Contradiction:
Manufacturing precisionVSUse of energy by moving object

Solution Approach 1:

The pixel processing is segmented into two categories: pixels processed with standard linear interpolation and pixels processed with high-precision non-linear filtering. This segmentation allows the system to maintain low power consumption for the majority of pixels while applying high precision conversion only where necessary for color accuracy.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

High precision color conversion is applied partially to only those pixels that require it, rather than to all pixels. The patent identifies pixels in gradient regions and applies the computationally intensive non-linear filtering only to these pixels, thereby reducing overall power consumption while maintaining color accuracy where needed.

Inventive Principle:
Principle #16Partial or excessive action

Data Source

PatentEP4496302B1Non-linear filtering for color space conversions
Publication Date: 2026.04.29 QUALCOMM INC
  • EP4496302B1 patent drawingFigure 1
  • EP4496302B1 patent drawingFigure 2
  • EP4496302B1 patent drawingFigure 3

AI summary

Aspects presented herein relate to methods and devices for graphics processing including an apparatus, e.g., a GPU. The apparatus may receive a plurality of pixels associated with a first color space including a plurality of first color channels, at least one first color channel of the plurality of first color channels being a first compressed channel. The apparatus may also decompress the at least one first color channel of the plurality of first color channels, the at least one first color channel being decompressed from the first compressed channel to a first decompressed channel. Further, the apparatus may perform a color space conversion of the first color space associated with the plurality of pixels, such that the plurality of first color channels is converted to a plurality of second color channels, the plurality of second color channels being associated with a second color space for the plurality of pixels.