Image Generating Device Balancing Viewpoint Responsiveness and Quality

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Image display technologies face challenges in balancing responsiveness to changes in the viewing point with maintaining high image quality, leading to impaired presence due to increased image processing load and latency.

Innovation Solution

An image generating device that includes a reference image data storage section, a viewing point information acquisition section, a projection section, and a pixel value determination section, which uses pre-acquired reference images to determine pixel values for the display image, allowing for efficient and accurate representation of objects based on predefined reference viewing points.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If complicated calculations are performed to raise resolution and heighten the sense of presence, then image quality is improved, but image processing load increases and responsiveness to viewing point changes deteriorates

Engineering Contradiction:
Improveimage resolutionVSAvoidresponsiveness to viewing point changes
Core Design Contradiction:
Measurement precisionVSSpeed

Solution Approach 1:

The patent pre-calculates and stores pixel values for multiple reference viewing points before actual display. When a viewing point changes, the system retrieves pre-computed pixel values from storage rather than performing real-time calculations, thereby maintaining high image resolution while achieving rapid responsiveness to viewing point changes.

Inventive Principle:
Principle #10Preliminary action

2Measurement precision

If real-time calculation is performed for high resolution display, then image quality is improved, but processing time increases causing latency

Engineering Contradiction:
Improvepixel value accuracyVSAvoidprocessing latency
Core Design Contradiction:
Measurement precisionVSLoss of time

Solution Approach 1:

The system performs pixel value calculations in advance for multiple reference viewing points and stores them in a storage unit. During actual display, pixel values are retrieved from storage based on the current viewing point, eliminating real-time calculation latency while maintaining high pixel value accuracy through pre-computed ray tracing or other sophisticated rendering methods.

Inventive Principle:
Principle #10Preliminary action

3Measurement precision

If sophisticated rendering methods like ray tracing are used, then image quality and sense of presence are improved, but computational complexity and processing load increase

Engineering Contradiction:
Improveimage realismVSAvoidcalculation complexity
Core Design Contradiction:
Measurement precisionVSDevice complexity

Solution Approach 1:

The patent applies sophisticated rendering methods such as ray tracing in advance to generate pixel values for multiple reference viewing points, which are then stored for rapid retrieval. This approach achieves high image realism through complex calculations while reducing real-time computational complexity by performing the heavy lifting beforehand and storing results for quick access during actual display operations.

Inventive Principle:
Principle #10Preliminary action

Data Source

PatentUS11120613B2Image generating device and method of generating image
Publication Date: 2021.09.14 SONY INTERACTIVE ENTERTAINMENT LLC
  • US11120613B2 patent drawing
  • US11120613B2 patent drawing
  • US11120613B2 patent drawing

AI summary

To achieve a balance between responsiveness of image display with respect to movement of a viewing point, and image quality. Reference viewing points are set in respect of a space containing an object to be displayed, and images of the space as viewed from these respective reference viewing points are created as reference images. Meanwhile, when pixel values of display images from a virtual camera are determined, reference images are selected wherein a point on the object represented by the pixel in question is represented as an image, and the values of these pixels are combined using a rule based on the positional relationship etc. of the reference viewing points with the virtual camera.