Stereoscopic Image Generation Using Pre-calculated Vector Maps

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing stereoscopic video technologies face challenges in generating high-quality images with reduced delay when the point of view is displaced, leading to potential distortion and decreased picture quality due to increased processing time and complex operations required for accurate parallax image generation.

Innovation Solution

An image generation apparatus and method that utilize original images from left and right viewpoints to calculate displacement vectors, generate a vector map, and filter reference vectors to determine pixel values for display images, allowing for efficient stereoscopic viewing without significant delay.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Manufacturing precision

If accurate parallax image generation processes (disposing object in virtual three-dimensional space, projecting with camera coordinate system changed) are applied to maintain image quality, then image quality is improved, but processing time increases and display delay occurs

Engineering Contradiction:
Improveimage qualityVSAvoidprocessing time
Core Design Contradiction:
Manufacturing precisionVSLoss of time

Solution Approach 1:

The patent pre-calculates and stores correspondence relationships between pixels in the left and right images before display. By preparing the vector map in advance that maps each pixel's position in one image to its corresponding position in the other image, the system avoids performing complex coordinate transformations in real-time during display, thus reducing processing delay while maintaining image quality

Inventive Principle:
Principle #10Preliminary action

Solution Approach 2:

The patent creates a vector map that copies and stores the spatial correspondence information between left and right images. Instead of repeatedly performing complex geometric transformations on the original images during display, the system uses this pre-generated vector map to quickly retrieve and apply correspondence relationships, significantly reducing processing time while preserving image accuracy

Inventive Principle:
Principle #26Copying

2Adaptability or versatility

If multiple operations are applied to initial parallax image data to enable point of view movement, then adaptability to different viewing positions is improved, but picture quality degradation increases

Engineering Contradiction:
Improvepoint of view movement capabilityVSAvoidpicture quality
Core Design Contradiction:
Adaptability or versatilityVSManufacturing precision

Solution Approach 1:

The patent extracts only the essential correspondence relationship information between left and right images into a vector map, separating this spatial mapping data from the actual image data. This allows the system to perform point of view movements by simply referencing the extracted correspondence information without applying multiple degrading operations to the original image data, thus maintaining picture quality while achieving adaptability

Inventive Principle:
Principle #2Taking out (Extraction)

Solution Approach 2:

The vector map serves as an intermediary data structure that stores the correspondence relationships between pixels in left and right images. Instead of directly manipulating the original image data through multiple transformation operations, the system uses this intermediary vector map to mediate the point of view movement process, reducing quality degradation while enabling adaptability to different viewing positions

Inventive Principle:
Principle #24Intermediary (Mediator)

Data Source

PatentEP4161070B1Image generation apparatus and image generation method
Publication Date: 2024.07.17 SONY INTERACTIVE ENTERTAINMENT LLC
  • EP4161070B1 patent drawingFigure 1
  • EP4161070B1 patent drawingFigure 2
  • EP4161070B1 patent drawingFigure 3

AI summary

A position and posture acquisition unit 250 of an image generation apparatus 200 acquires position information relating to a point of view of a user. A viewscreen setting unit 252 sets a viewscreen. An original image operation unit 254 calculates a correction amount for a pixel from parallax values of pixels of an original image and an amount of movement of the point of view such that an object looks fixed. Then, an image reference vector map in which an image reference vector that refers to a position before correction from coordinates of each pixel after correction is stored for each pixel is generated. A display image generation unit 268 specifies an image reference vector corresponding to each pixel of the viewscreen and generates, based on the image reference vectors, a color value by filtering pixel values of reference destinations in the original image. An outputting unit 270 outputs a display image.