Volumetric Video Viewpath Forecasting for Distortion-Free Inserts

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing methods fail to adequately insert digital information and advertisements into volumetric video without angular distortion and to remove obstructing objects, ensuring seamless user experience across varying viewpoints.

Innovation Solution

A method that captures multiple fields of view, analyzes excitement levels, forecasts a virtual field of view path, and composites data from cameras to insert images while managing angular offset and object transparency.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Ease of manufacture

If digital content is inserted into volumetric video from a fixed camera viewpoint, then the insertion process is simple, but angular distortion occurs when users view from different angles

Engineering Contradiction:
Improveinsertion process simplicityVSAvoidangular distortion
Core Design Contradiction:
Ease of manufactureVSManufacturing precision

Solution Approach 1:

The system dynamically adjusts the field of view parameters and camera viewpoints based on the user's viewing angle. The virtual camera's field of view is modified in real-time to compensate for angular offsets, ensuring that inserted digital content remains undistorted regardless of the user's viewing position. This dynamic adaptation resolves the contradiction by making the insertion process angle-independent.

Inventive Principle:
Principle #15Dynamics

Solution Approach 2:

The system changes key parameters including the virtual camera's field of view angle, focal length, and position to maintain proper perspective for inserted content. By adjusting these parameters based on the user's viewing angle, the system eliminates angular distortion while keeping the insertion process straightforward.

Inventive Principle:
Principle #35Parameter changes

2Adaptability or versatility

If multiple cameras are used to capture volumetric video from different viewpoints, then user viewing flexibility is improved, but system complexity increases

Engineering Contradiction:
Improveuser viewing flexibilityVSAvoidsystem complexity
Core Design Contradiction:
Adaptability or versatilityVSDevice complexity

Solution Approach 1:

Instead of using multiple physical cameras to capture all possible viewpoints, the system uses a single physical camera to capture volumetric data and then creates multiple virtual camera viewpoints through computational rendering. This copying approach provides user viewing flexibility while avoiding the complexity of deploying and synchronizing multiple physical cameras.

Inventive Principle:
Principle #26Copying

Solution Approach 2:

The system transitions from a two-dimensional camera plane to a three-dimensional volumetric space, allowing virtual cameras to be positioned anywhere within the captured volume. This dimensional change enables flexible viewpoint selection without adding physical camera complexity.

Inventive Principle:
Principle #17Another dimension (Dimensionality change)

3Ease of manufacture

If advertisements are placed in fixed locations within the scene, then insertion is straightforward, but visibility is limited for users viewing from different angles

Engineering Contradiction:
Improveinsertion straightforwardnessVSAvoidadvertisement visibility
Core Design Contradiction:
Ease of manufactureVSReliability

Solution Approach 1:

The system dynamically repositions and re-orientates inserted advertisements based on the user's current viewing angle. As the user moves through different viewpoints, the advertisements automatically adjust their position and orientation within the volumetric scene to remain visible and properly oriented, ensuring consistent visibility while keeping insertion straightforward.

Inventive Principle:
Principle #15Dynamics

Data Source

PatentUS12537909B2Virtual field of view adjustment in live volumetric video
Publication Date: 2026.01.27 INTERNATIONAL BUSINESS MACHINE CORPORATION
  • US12537909B2 patent drawing
  • US12537909B2 patent drawing
  • US12537909B2 patent drawing

AI summary

Video of a plurality of fields of view of a scene is captured, each field of view comprising data of the scene from a different vantage point. An excitement level is determined by analyzing a portion of the captured video. Using the excitement level, a time series of future excitement levels is forecast. Using the time series of future excitement levels, a virtual field of view path of the scene is forecast. An insert image is determined to be included in the virtual field of view path. Captured data from at least two video cameras in the plurality of video cameras is composited into a virtual field of view of the scene. A rendering of the insert image is inserted into the virtual field of view.