Occluded Object Area Determination via Feature Reliability Assessment

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Existing methods for detecting a person in an image struggle with accuracy when parts of the object are occluded, leading to deterioration in the extraction of image features and determination of object areas.

Innovation Solution

An image processing apparatus is designed with a first detection unit to identify feature points corresponding to object parts, an acquisition unit to assess the reliability of these feature points, a second detection unit to correct feature points with low reliability, and a determination unit to accurately determine the object area based on these features.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If feature points are detected from occluded peripheral areas, then the object area can be determined, but the accuracy of feature extraction deteriorates

Engineering Contradiction:

Improvefeature extraction accuracyVSAvoidocclusion

Core Design Contradiction:

Measurement precisionVSObject-affected harmful factors

Solution Approach 1:

The patent converts the harmful effect of occlusion into a beneficial filtering mechanism. By assessing the reliability of each feature point and identifying those with low reliability (caused by occlusion), the system selectively excludes these problematic feature points from object area determination. This transforms the harmful occlusion effect into a useful criterion for improving overall measurement accuracy.

Inventive Principle:

Principle #22Blessing in disguise (Convert harm into benefit)

Solution Approach 2:

The patent applies different quality standards to different feature points based on their individual reliability assessments. Instead of uniformly treating all feature points, the system evaluates each point's reliability separately and selectively uses only high-reliability feature points for object area determination. This local quality approach ensures that occluded or low-quality regions do not degrade the overall measurement accuracy.

Inventive Principle:

Principle #3Local quality

2Measurement precision

If a circumscribed rectangle is drawn around a feature point group to determine object area, then the object area can be determined, but accuracy deteriorates when parts of the object are occluded

Engineering Contradiction:

Improveobject area determination accuracyVSAvoiddetermination reliability under occlusion

Core Design Contradiction:

Measurement precisionVSReliability

Solution Approach 1:

The patent segments the object area determination process into two distinct stages: first, detecting multiple feature points across the object; second, assessing the reliability of each feature point and selectively using only high-reliability points for the final circumscribed rectangle calculation. This segmentation allows the system to identify and exclude feature points affected by occlusion, thereby improving the accuracy of the final object area determination.

Inventive Principle:

Principle #1Segmentation

Solution Approach 2:

The patent introduces a feedback mechanism where the reliability of each feature point is assessed and used to inform the subsequent object area determination. The system feeds back the reliability information to selectively include or exclude feature points, creating a closed-loop process that continuously improves determination accuracy by eliminating the influence of occluded regions.

Inventive Principle:

Principle #23Feedback

3Measurement precision

If feature points from occluded areas are used for re-identification, then identification can be performed, but the accuracy of person identification deteriorates

Engineering Contradiction:

Improveperson identification accuracyVSAvoididentification information quality

Core Design Contradiction:

Measurement precisionVSLoss of information

Solution Approach 1:

The patent converts the information loss caused by occlusion into a beneficial filtering opportunity. By assessing feature point reliability and identifying those with low reliability (indicating occlusion), the system selectively excludes these degraded information sources from re-identification. This transforms the harmful information loss into a useful criterion for selecting only high-quality identification features.

Inventive Principle:

Principle #22Blessing in disguise (Convert harm into benefit)

Solution Approach 2:

The patent applies local quality assessment to each feature point individually, evaluating its reliability based on occlusion conditions. Instead of uniformly processing all feature points for re-identification, the system selectively uses only high-reliability feature points, ensuring that occluded or low-quality regions do not degrade the overall identification accuracy.

Inventive Principle:

Principle #3Local quality

Data Source

PatentUS12217528B2Image processing apparatus for determining an object area where part of the object is occluded

Publication Date: 2025.02.04 CANON KK

AI summary

An image processing apparatus includes a first detection unit configured to detect, from an image in which an object including a plurality of parts is captured, first feature points corresponding to the parts of the object, an acquisition unit configured to acquire a reliability indicating a likelihood that a position indicated by a feature point is a part corresponding to the feature point for each of the first feature points detected by the first detection unit, a second detection unit configured to detect a second feature point based on some of the first feature points for a part corresponding to a first feature point with the low reliability, and a determination unit configured to determine an area including the object based on some of the first feature points and the second feature point.