Virtual Viewpoint Data Subject Positioning
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing techniques for generating virtual viewpoint video struggle to accurately arrange subjects at desired positions, as determining appropriate virtual viewpoints is complex due to limitations in data generation methods, particularly when capturing multiple subjects in the same image-capturing area.
Innovation Solution
An information processing apparatus that acquires parameters representing the position and line-of-sight direction of a virtual viewpoint, specifying the position and orientation of subjects, and outputs virtual viewpoint data associated with these parameters to generate desired virtual viewpoint videos, utilizing virtual camera path data that includes three-dimensional position and attitude parameters of a virtual camera.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Manufacturing precision
If a virtual viewpoint is determined using existing methods, then virtual viewpoint video can be generated, but the subject cannot be accurately arranged at desired positions in the video
Solution Approach 1:
The patent applies preliminary action by pre-defining the relationship between virtual viewpoint parameters and subject position information. The system预先 establishes the mapping between camera parameters (position, orientation) and subject arrangement data, enabling accurate subject placement at desired positions in the virtual viewpoint video without requiring complex real-time calculations.
2Adaptability or versatility
If multiple subjects are captured in the same image-capturing area, then comprehensive video content is achieved, but determining appropriate virtual viewpoints becomes difficult
Solution Approach 1:
The patent applies segmentation by separating the handling of multiple subjects through individual position and orientation parameters for each subject. The system divides the complex task of managing multiple subjects into independent parameter sets (position, orientation) for each subject, allowing the virtual viewpoint system to accurately track and render each subject independently while maintaining overall scene coherence.
3Ease of manufacture
If virtual viewpoint data is generated without subject position information, then data generation is simplified, but the subject arrangement at desired positions is not achieved
Solution Approach 1:
The patent introduces an intermediary element - subject position and orientation parameters - that mediates between the virtual viewpoint parameters and the final subject arrangement. These intermediate parameters serve as a bridge, allowing the system to maintain simplified data generation processes while ensuring accurate subject positioning through the additional position and orientation information.
Data Source
AI summary
A virtual camera path data processing apparatus includes one or more memories storing instructions and one or more processors executing the instructions to acquire a parameter that represents a position of a virtual viewpoint and a line-of-sight direction from the virtual viewpoint, the parameter corresponding to a frame included in virtual viewpoint video generated based on a plurality of captured images of a subject, the plurality of captured images being obtained by a plurality of image capturing devices, specify a position and an orientation of the subject in the virtual viewpoint video, and output virtual viewpoint data in which the acquired parameter is associated with predetermined information based on the specified position and the specified orientation of the subject.