Object Tracking Device for Occlusion Management

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Existing object tracking methods fail to accurately track objects when mutual occlusion occurs among independently moving objects, leading to decreased prediction accuracy and difficulties in maintaining accurate tracking.

Innovation Solution

An object tracking system that includes object detection, tracking, grouping determination, separation determination, and tracking correction mechanisms to identify and correct for occlusions among objects, ensuring accurate tracking even during mutual occlusion.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If object tracking is performed using conventional methods, then tracking can be maintained under normal conditions, but tracking accuracy deteriorates when mutual occlusion occurs among objects

Engineering Contradiction:

Improvetracking accuracyVSAvoidtracking reliability under occlusion

Core Design Contradiction:

Measurement precisionVSReliability

Solution Approach 1:

The tracking system segments objects into groups based on their spatial proximity and occlusion relationships. By dividing the tracking problem into individual object tracking and group-level occlusion management, the system can apply different strategies for visible and occluded objects, maintaining overall tracking accuracy even when mutual occlusion occurs

Inventive Principle:

Principle #1Segmentation

Solution Approach 2:

The system dynamically changes tracking parameters based on occlusion detection. When mutual occlusion is detected within a group, the system adjusts prediction models and tracking confidence thresholds specifically for occluded objects, allowing continuous tracking despite reduced visibility

Inventive Principle:

Principle #35Parameter changes

2Productivity

If tracking target is shifted to occluding object when tracking becomes abnormal, then tracking can continue, but accuracy of tracking the original occluded object deteriorates

Engineering Contradiction:

Improvetracking continuityVSAvoidoriginal object tracking accuracy

Core Design Contradiction:

ProductivityVSMeasurement precision

Solution Approach 1:

The system performs preliminary grouping determination to identify potential occlusion relationships before tracking abnormalities occur. By pre-establishing which objects are likely to occlude each other based on their trajectories and spatial relationships, the system can maintain tracking of the original occluded object using prediction models even when the object becomes temporarily invisible

Inventive Principle:

Principle #10Preliminary action

Solution Approach 2:

The system implements feedback mechanisms where grouping determination results continuously inform tracking corrections. When objects are identified as being in a mutual occlusion relationship, the tracking system receives feedback to maintain separate tracking states for both objects, using group-level information to correct individual object trajectories and prevent target switching

Inventive Principle:

Principle #23Feedback

3Duration of action of moving object

If temporal prediction model is incorporated in tracking process, then tracking can be maintained during occlusion, but prediction accuracy deteriorates when mutual occlusion occurs

Engineering Contradiction:

Improvetracking duration during occlusionVSAvoidposition prediction accuracy

Core Design Contradiction:

Duration of action of moving objectVSMeasurement precision

Solution Approach 1:

The system adds a group-level dimension to the tracking process by introducing grouping determination that operates alongside individual object tracking. This additional dimension allows the system to consider occlusion relationships and spatial proximity when making predictions, correcting trajectory estimates based on group behavior patterns even when individual object visibility is poor

Inventive Principle:

Principle #17Another dimension (Dimensionality change)

Data Source

PatentUS10867394B2Object tracking device, object tracking method, and recording medium

Publication Date: 2020.12.15 NEC CORP

AI summary

An object tracking means tracks objects in a video image. An object consolidation determination means generates a “container” which includes a plurality of objects located in proximity. When an object separation determination means determines that an object is released from the “container”, the object tracking means restarts tracking the object released from the “container”.