Video-Based Pedestrian Risk Prediction to Reduce False Alarms

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing traffic accident warning systems are oversensitive and generate frequent false alarms due to relying solely on past and current object locations, and pedestrian behavior prediction methods based on image data have insufficient accuracy, making proactive accident prevention challenging.

Innovation Solution

A system and method for predicting pedestrian safety information using video analysis that includes pedestrian trajectory prediction, behavior prediction, and risk assessment, employing attention mechanisms and neural networks to generate proactive warnings.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Device complexity

If traffic accident warning systems use only past and current object locations, then the system is simple to implement, but it generates frequent false alarms and reduces reliability

Engineering Contradiction:
Improvesystem complexityVSAvoidwarning reliability
Core Design Contradiction:
Device complexityVSReliability

Solution Approach 1:

The system performs preliminary trajectory prediction by analyzing future potential paths of pedestrians and vehicles before accidents occur. This proactive approach allows the system to anticipate dangerous situations and generate warnings in advance, reducing false alarms by predicting actual collision risks rather than reacting to past positions alone

Inventive Principle:
Principle #10Preliminary action

Solution Approach 2:

The system incorporates feedback mechanisms by continuously monitoring video feeds, comparing predicted trajectories with actual movements, and adjusting warning generation based on real-time observations. This feedback loop enables the system to learn from past predictions and improve accuracy, reducing false alarms while maintaining high reliability

Inventive Principle:
Principle #23Feedback

2Device complexity

If pedestrian behavior prediction uses only image data, then the system is simpler, but prediction accuracy is insufficient

Engineering Contradiction:
Improvedata processing complexityVSAvoidprediction accuracy
Core Design Contradiction:
Device complexityVSMeasurement precision

Solution Approach 1:

The system merges multiple data sources including image data, video sequences, and trajectory predictions into a unified prediction framework. By combining visual information from cameras with motion trajectory analysis, the system achieves higher prediction accuracy while managing complexity through integrated processing

Inventive Principle:
Principle #5Merging (Combining)

Data Source

PatentEP4498335B1System and method for predicting pedestrian safety information based on video
Publication Date: 2026.01.28 ELECTRONICS & TELECOMM RES INST
  • EP4498335B1 patent drawingFigure 1
  • EP4498335B1 patent drawingFigure 2
  • EP4498335B1 patent drawingFigure 3

AI summary

A method and apparatus for predicting pedestrian safety information based on video is disclosed. Pedestrian trajectory prediction data based on video input is first generated. Then, pedestrian behavior prediction data based on the video input is generated. A potential risk to pedestrian safety is estimated based on the pedestrian trajectory prediction data, the pedestrian behavior prediction data, and a surface classification data in the video data.