Image Rendering with Flexible Scale Rasterization Metadata

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing technologies fail to efficiently address the rendering of immersive content, such as content, and the need for efficient rendering of high-quality video and video game content, particularly in VR, due to hardware limitations in processing power.

Innovation Solution

Implement flexible scale rasterization (FSR) to vary rasterization rates based on image content and viewer gaze, using metadata to optimize texture rendering without modifying shaders, and support multiple FSR settings for different image regions.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Manufacturing precision

If high-quality rendering is applied to entire image, then content quality is improved, but processing power requirement increases

Engineering Contradiction:
Improveimage qualityVSAvoidprocessing power
Core Design Contradiction:
Manufacturing precisionVSPower

Solution Approach 1:

The patent applies different rendering qualities to different regions of the image based on the flexible scale rasterization scheme. High-quality rendering is applied only to regions that require it (such as foveal regions in VR), while other regions use lower quality rendering. This resolves the contradiction by making image quality non-uniform, matching the actual visual requirements of different areas and reducing overall processing power needs while maintaining perceived quality.

Inventive Principle:
Principle #3Local quality

2Manufacturing precision

If uniform high resolution is used across the image, then content quality is improved, but processing burden increases

Engineering Contradiction:
Improveimage resolutionVSAvoidprocessing efficiency
Core Design Contradiction:
Manufacturing precisionVSProductivity

Solution Approach 1:

The patent segments the image into multiple regions with different resolution requirements based on the flexible scale rasterization scheme. Each region is rendered at an appropriate resolution level, with higher resolution allocated to visually important areas and lower resolution to less critical areas. This segmentation approach improves processing efficiency by avoiding the unnecessary rendering of entire high-resolution images while maintaining quality where needed.

Inventive Principle:
Principle #1Segmentation

3Manufacturing precision

If processing power is increased to improve content quality, then content quality is improved, but hardware cost increases

Engineering Contradiction:
Improvecontent qualityVSAvoidhardware capability
Core Design Contradiction:
Manufacturing precisionVSDevice complexity

Solution Approach 1:

The patent changes the parameter of image resolution from a uniform value to a spatially varying value based on the flexible scale rasterization scheme. By adjusting the resolution parameter locally across different image regions rather than using a single high resolution throughout, the system achieves high content quality in critical areas while reducing the overall processing power and hardware capability requirements.

Inventive Principle:
Principle #35Parameter changes

Data Source

PatentEP4224407B1Image processing system and method
Publication Date: 2025.12.17 SONY INTERACTIVE ENTERTAINMENT LLC
  • EP4224407B1 patent drawingFigure 1~2
  • EP4224407B1 patent drawingFigure 3~4
  • EP4224407B1 patent drawingFigure 5~6

AI summary

An image processing system operable to generate one or more images for display, the system comprising a rendering unit operable to render one or more portions of an image using a plurality of flexible scale rasterization, FSR, schemes, a metadata generation unit operable to generate metadata indicating one or more properties of each of the FSR schemes used to render a particular image portion, a shader compiler operable to read the metadata, and to modify a shader to enable an image portion sampling to be performed that includes a corresponding FSR resolve where the metadata indicates that the particular image portion has been rendered with FSR, and an image generation unit operable to generate an image for display using at least one modified shader, the generated image including one or more of the plurality of image portions.