Virtual Viewpoint Image Generation for Panoramic Player Detail
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for generating virtual viewpoint images struggle to provide clear visualization of player positions and facial expressions/sights in sports broadcasting, as players appear small and details are difficult to discern.
Innovation Solution
An image processing apparatus that generates virtual viewpoint images by using multiple cameras to capture an image capturing region, separates foreground and background, and deforms foreground models to enhance visibility, allowing for panoramic views.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Area of stationary object
If a virtual viewpoint image including the entire field is used to indicate player positions, then the viewer can check player positions and formation on the field, but the player is displayed small and it is difficult to check facial expressions and lines of sight
Solution Approach 1:
The image processing apparatus segments the virtual viewpoint image into multiple regions: a first region displaying the entire field for position/formation overview, and a second region displaying an enlarged portion of a specific player for detailed facial expression and line of sight information. This segmentation allows simultaneous presentation of both global context and local details without compromising either.
Solution Approach 2:
The patent transitions from a single two-dimensional display to a multi-region two-dimensional display that effectively creates an additional dimension of information presentation. By dividing the display area into multiple functional regions (first region for overview, second region for detail), it enables viewers to access both broad spatial context and fine-grained player details within the same visual field.
2Area of stationary object
If a graphic representing the entire field is used to indicate player positions, then the viewer can check player positions and formation, but the viewer cannot check facial expression and line of sight of the player
Solution Approach 1:
The display is segmented into a first region showing the entire field with player position graphics and a second region showing an enlarged view of a specific player's facial features and orientation. This allows the system to maintain comprehensive field coverage while simultaneously providing access to detailed player expressions and line of sight information that would be invisible in a full-field view.
Solution Approach 2:
The system creates a copied or replicated view of a specific player from the full-field graphic, displaying it in an enlarged format in the second region. This copy contains detailed information about facial expressions and line of sight that is not visible in the original full-field representation, allowing viewers to examine player details without losing the context of the overall field situation.
3Area of stationary object
If the player is displayed small on a virtual viewpoint image, then the entire field can be shown, but it is difficult for the viewer to check facial expressions and lines of sight
Solution Approach 1:
The display area is segmented into a first region for showing the entire field with small-scale player representations and a second region for showing an enlarged, high-detail view of a specific player. This segmentation resolves the contradiction by allowing small player icons in the first region to maintain field coverage while the enlarged player in the second region provides the measurement precision needed for facial expression and line of sight analysis.
Solution Approach 2:
Different regions of the display are assigned different quality characteristics: the first region uses small-scale representations suitable for showing overall field coverage and player positions, while the second region uses high-resolution enlarged imagery specifically optimized for displaying facial expressions and line of sight details. Each region's quality is optimized for its specific purpose.
Data Source
AI summary
An image processing apparatus: obtain information on a virtual viewpoint to generate a virtual viewpoint image, the virtual viewpoint image being an image of an object viewed from the virtual viewpoint, the object included in an image capturing region of an image capturing apparatus; obtain enlarged three-dimensional shape data representing the object in which at least a part of regions is greater than that of normal three-dimensional shape data generated based on images obtained by image capturing by the image capturing apparatus; and generate the virtual viewpoint image based on the enlarged three-dimensional shape data in a case where a visual field represented by the information is a panoramic visual field.