Video Left-Object Detection with Person Presence Verification

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Existing technologies inaccurately detect stationary objects as left objects, such as luggage left by a seated person, and fail to identify objects left by individuals who do not move, leading to false detections.

Innovation Solution

An information processing apparatus that includes a stationary object detection unit, a person detection unit, and a process execution unit to analyze video frames before and after detecting a stationary object, determining if a person is present in its vicinity to accurately identify left objects.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If an object being continuously stationary is detected as a left object, then the detection sensitivity is improved, but the false detection rate increases

Engineering Contradiction:

Improvedetection sensitivityVSAvoidfalse detection rate

Core Design Contradiction:

Measurement precisionVSReliability

Solution Approach 1:

The detection process is segmented into multiple independent analysis dimensions: stationary object detection, person detection, and relationship analysis. By dividing the detection task into separate modules that analyze different aspects (object status, person presence, spatial relationship), the system achieves both high sensitivity in detecting left objects and high reliability in reducing false alarms through comprehensive verification

Inventive Principle:

Principle #1Segmentation

Solution Approach 2:

Person detection results serve as an intermediary verification mechanism between stationary object detection and final left object determination. The presence or absence of a person in vicinity of the stationary object acts as a mediating factor that confirms whether the stationary object is truly a left object or merely a temporarily placed item, thereby reducing false detections

Inventive Principle:

Principle #24Intermediary (Mediator)

2Measurement precision

If person detection is performed for each video frame, then the accuracy of identifying left objects is improved, but the processing time increases

Engineering Contradiction:

Improveidentification accuracyVSAvoidprocessing time

Core Design Contradiction:

Measurement precisionVSLoss of time

Solution Approach 1:

The system performs preliminary stationary object detection on video frames before conducting person detection. By first identifying stationary objects and then selectively performing person detection only in relation to these stationary objects, the system reduces overall processing time while maintaining high identification accuracy for left objects

Inventive Principle:

Principle #10Preliminary action

Solution Approach 2:

Instead of performing full person detection on all objects in all frames, the system applies person detection selectively only to stationary objects that are candidates for left objects. This partial action approach reduces processing time significantly while maintaining sufficient accuracy for the specific task of identifying left objects

Inventive Principle:

Principle #16Partial or excessive action

Data Source

PatentUS12423985B2Information processing apparatus, control method, and program

Publication Date: 2025.09.23 NEC CORP

AI summary

An information processing apparatus (2000) detects a stationary object from video data (12). In addition, the information processing apparatus (2000) executes person detection process of detecting a person in vicinity of an object (target object) detected as the stationary object for each of a plurality of video frames (14) which includes the target object. Furthermore, the information processing apparatus (2000) executes a predetermined process by comparing results of the person detection process for each of the plurality of video frames (14).