3D Object-Based Camera Feed Selection in Multi-Camera Video

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Existing virtual viewpoint video generation systems struggle with efficiently switching to real camera videos of specific objects, especially when multiple performers are present, as users must manually search through numerous camera feeds, which is cumbersome and inefficient.

Innovation Solution

A video processing apparatus that utilizes three-dimensional shape information to identify and display candidate camera feeds capturing a designated object, reducing the number of feeds the user needs to review by specifying the image capturing apparatus based on bounding boxes defined by the object's three-dimensional shape.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Area of stationary object

If a large number of cameras are installed to expand the image capturing range and improve image quality, then the image capturing capability and coverage are improved, but the complexity of camera management increases and it becomes difficult for users to confirm and switch to specific camera feeds

Engineering Contradiction:

Improveimage capturing rangeVSAvoidcamera feed selection

Core Design Contradiction:

Area of stationary objectVSEase of operation

Solution Approach 1:

The patent introduces an object-based intermediary system that mediates between the user and the large number of cameras. Instead of directly managing camera feeds, the system uses detected objects (performers) as intermediaries. Users select objects of interest, and the system automatically identifies and presents relevant camera feeds capturing those objects, thus simplifying the complex camera management task.

Inventive Principle:

Principle #24Intermediary (Mediator)

Solution Approach 2:

The patent replaces the manual mechanical process of browsing through numerous camera feeds with an automated computer vision system. The system uses object detection algorithms to automatically identify performers in the scene and determine which cameras are capturing them, substituting automated image processing for manual camera selection.

Inventive Principle:

Principle #28Mechanics substitution (Replace mechanical system)

2Measurement precision

If users manually check videos from multiple real cameras to find a specific performer, then the ability to locate specific content is achieved, but the time and operational burden required increases significantly

Engineering Contradiction:

Improveobject identification accuracyVSAvoidcamera switching time

Core Design Contradiction:

Measurement precisionVSLoss of time

Solution Approach 1:

The system performs preliminary object detection and camera identification in advance. Instead of waiting for users to manually search through camera feeds, the system proactively detects objects in the scene, identifies which cameras are capturing those objects, and prepares the information for immediate user selection, significantly reducing the time required to switch to specific camera feeds.

Inventive Principle:

Principle #10Preliminary action

Solution Approach 2:

The system provides feedback to users by displaying information about which cameras are capturing selected objects. This feedback mechanism allows users to make informed decisions about which camera feeds to switch to, based on real-time information about object-camera relationships, thereby reducing the time and effort needed to locate desired content.

Inventive Principle:

Principle #23Feedback

3Reliability

If the system displays only videos from cameras used for virtual viewpoint generation, then the virtual viewpoint generation function is supported, but the ability to switch to real camera videos showing specific objects not used in virtual viewpoint generation is lost

Engineering Contradiction:

Improvevirtual viewpoint generationVSAvoidreal camera switching flexibility

Core Design Contradiction:

ReliabilityVSAdaptability or versatility

Solution Approach 1:

The patent implements a universal object detection and camera identification system that serves multiple functions. The same system that identifies cameras for virtual viewpoint generation is also used to identify cameras capturing any object of interest, regardless of whether that object is used in virtual viewpoint generation. This multi-functional approach allows the system to support both virtual viewpoint generation and flexible real camera switching to specific objects.

Inventive Principle:

Principle #6Universality (Multi-functionality)

Data Source

PatentUS12556672B2Video processing apparatus for designating an object on a predetermined video and control method of the same, and storage medium

Publication Date: 2026.02.17 CANON KK

AI summary

A video processing apparatus accepts, from a user, designation of an object on a predetermined video captured by a first real image capturing apparatus, specifies, from a plurality of second real image capturing apparatuses, an image capturing apparatus that captured a video including the designated object in an image capturing range based on three dimensional shape information corresponding to the designated object. The designated object exists in an image capturing area captured by the plurality of second real image capturing apparatuses. The video processing apparatus outputs display information including presentation of the specified image capturing apparatus.