3D Multi-View Object Tracking With Periodic Error Correction

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing object tracking technologies struggle with errors and inaccuracies, particularly in crowded or occluded environments such as ball sports, leading to erroneous tracking results due to players or structures obstructing the view.

Innovation Solution

An information processing apparatus that combines multi-viewpoint frame analysis with both efficient tracking and identification methods, using identifiers to track object positions accurately by integrating tracking unit and identification unit processes, and correcting errors through complementary updates.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Productivity

If object tracking uses past tracking results and current state information to estimate current position, then tracking efficiency is improved, but tracking accuracy deteriorates in occluded or crowded environments

Engineering Contradiction:
Improvetracking efficiencyVSAvoidtracking accuracy
Core Design Contradiction:
ProductivityVSMeasurement precision

Solution Approach 1:

The patent implements feedback by periodically identifying the object from captured images and comparing the identified position with the tracked position. When discrepancies exceed a threshold, the system corrects the tracking result using the identified position as feedback, thereby maintaining accuracy while preserving tracking efficiency.

Inventive Principle:
Principle #23Feedback

Solution Approach 2:

The system performs preliminary identification of the object from captured images at specific intervals before final tracking decisions are made. This preliminary action provides reference information that helps correct potential tracking errors before they accumulate, balancing efficiency with accuracy.

Inventive Principle:
Principle #10Preliminary action

2Use of energy by moving object

If tracking continuously uses past results to estimate current position, then computational load is reduced, but error accumulation increases leading to erroneous tracking

Engineering Contradiction:
Improvecomputational loadVSAvoidtracking reliability
Core Design Contradiction:
Use of energy by moving objectVSReliability

Solution Approach 1:

The patent employs periodic action by performing full object identification from captured images at predetermined intervals rather than continuously. This periodic correction identifies and corrects accumulated tracking errors while maintaining low computational load during intermediate periods, thus preserving both efficiency and reliability.

Inventive Principle:
Principle #19Periodic action

Solution Approach 2:

The system performs self-correction by automatically detecting tracking errors through periodic identification and adjusting its own tracking results. The identification unit serves the tracking process by providing corrective feedback when needed, enabling the system to maintain reliability without continuous external intervention or heavy computational processing.

Inventive Principle:
Principle #25Self-service

3Measurement precision

If full identification processing is performed continuously to ensure accuracy, then tracking precision is improved, but computational complexity increases

Engineering Contradiction:
Improvetracking precisionVSAvoidcomputational complexity
Core Design Contradiction:
Measurement precisionVSDevice complexity

Solution Approach 1:

The patent reduces computational complexity by performing full identification processing only at predetermined intervals rather than continuously. During periodic intervals, the system executes comprehensive identification to correct tracking drift, while between intervals it relies on lighter-weight tracking estimation, thus achieving acceptable precision with reduced computational burden.

Inventive Principle:
Principle #19Periodic action

Solution Approach 2:

The system applies partial action by performing full identification processing only when necessary (at predetermined intervals or when error thresholds are exceeded) rather than continuously. This selective application of full identification maintains tracking precision where needed while avoiding unnecessary computational complexity during stable tracking periods.

Inventive Principle:
Principle #16Partial or excessive action

Data Source

PatentEP4697264A1Object tracking
Publication Date: 2026.02.18 CANON KK
  • EP4697264A1 patent drawingFigure 1
  • EP4697264A1 patent drawingFigure 2A~2B
  • EP4697264A1 patent drawingFigure 3

AI summary

An information processing apparatus(100) according to the present disclosure obtains a plurality of captured images obtained by capturing images of an object from different directions(S301), estimates a three-dimensional shape of the object using the plurality of captured images(S302), performs tracking processing of the object to track the object by estimating an identifier and a position of the object using the three-dimensional shape(S303), performs identification processing of the object to identify the object using at least some of the plurality of captured images(S311) and estimate an identifier and a position of the object, detects an error in the result of the tracking processing based on the result of the tracking processing and the result of the identification processing, and complements tracking data indicating the result of the tracking processing using the result of the identification processing in a case where an error in the tracking processing is detected(S315).