Video Behavior Prediction Using Person-Object Relationship Analysis

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing behavior recognition technologies are limited to recognizing behaviors that have already occurred, making it difficult to take timely countermeasures, and fail to predict potential future situations that may require intervention.

Innovation Solution

An information processing system that combines behavior analysis with context sensing to identify relationships and predict future behaviors by analyzing video data using machine learning models for skeleton and facial expression recognition, enabling the detection of potential future events such as shoplifting, illness, or crime.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If behavior recognition technology is used to recognize current or past behaviors, then behavior detection capability is improved, but the ability to predict future behaviors requiring countermeasures deteriorates

Engineering Contradiction:
Improvebehavior detection capabilityVSAvoidresponse time for countermeasures
Core Design Contradiction:
Measurement precisionVSLoss of time

Solution Approach 1:

The system performs preliminary analysis of behavior patterns, relationships between persons and objects, and contextual factors to predict future behaviors before they occur. This allows countermeasures to be prepared in advance, transforming reactive response into proactive prevention and eliminating the time loss between detection and intervention.

Inventive Principle:
Principle #10Preliminary action

2Device complexity

If simple behavior recognition is implemented, then system complexity is reduced, but the ability to identify situations requiring countermeasures deteriorates

Engineering Contradiction:
Improvesystem complexityVSAvoidaccuracy in identifying situations requiring countermeasures
Core Design Contradiction:
Device complexityVSReliability

Solution Approach 1:

The system segments the behavior analysis into distinct components: behavior pattern recognition, relationship identification (between persons and objects), and contextual factor analysis. Each component processes specific aspects independently, then integrates results to predict future behaviors. This segmentation maintains manageable system complexity while achieving comprehensive and reliable situation identification.

Inventive Principle:
Principle #1Segmentation

Data Source

PatentEP4207098B1Information processing program, information processing method, and information processing apparatus
Publication Date: 2026.03.25 FUJITSU LTD
  • EP4207098B1 patent drawingFigure 1
  • EP4207098B1 patent drawingFigure 2
  • EP4207098B1 patent drawingFigure 3

AI summary

An information processing apparatus acquires (51) video data that includes target objects including a person and an object, first identifies (52) a relationship between the target objects in the acquired video data, by inputting the acquired video data to a first machine learning model, second identifies (53) a behavior of the person in the video data by using a feature value of the person included in the acquired video data, and predicts (54) one of a future behavior and a future state of the person by comparing the identified behavior of the person and the identified relationship with a behavior prediction rule that is set in advance.