Dual Tolerance Object Detection for Image Processing Accuracy

Solution Overview

Problem

Conventional image processing methods face issues with inappropriate processing due to high or low detection reliability levels, leading to undetected or incorrectly detected objects, which results in unintended image processing.

Innovation Solution

An apparatus and method that utilize two detection units with different tolerance levels to detect objects, allowing selection based on the specific processing requirements, thereby minimizing inappropriate image processing by adjusting the detection criteria according to the processing tasks.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If the detection result reliability level is set high, then detection accuracy is improved, but detection coverage deteriorates causing undetected errors

Engineering Contradiction:

Improvedetection accuracyVSAvoiddetection coverage

Core Design Contradiction:

Measurement precisionVSReliability

Solution Approach 1:

The detection unit is divided into a first detection unit with narrow detection tolerance for high-accuracy processing and a second detection unit with wide detection tolerance for comprehensive coverage. This segmentation allows each detection unit to specialize in different aspects of object detection, resolving the contradiction between accuracy and coverage by having both units operate simultaneously on the same image data.

Inventive Principle:

Principle #1Segmentation

2Reliability

If the detection result reliability level is set low, then detection coverage is improved, but detection accuracy deteriorates causing detection errors

Engineering Contradiction:

Improvedetection coverageVSAvoiddetection accuracy

Core Design Contradiction:

ReliabilityVSMeasurement precision

Solution Approach 1:

The detection unit is divided into a first detection unit with narrow detection tolerance for high-accuracy processing and a second detection unit with wide detection tolerance for comprehensive coverage. This segmentation allows each detection unit to specialize in different aspects of object detection, resolving the contradiction between accuracy and coverage by having both units operate simultaneously on the same image data.

Inventive Principle:

Principle #1Segmentation

3Device complexity

If a single detection unit is used, then device complexity is reduced, but processing adaptability deteriorates

Engineering Contradiction:

Improvedetection system structureVSAvoidprocessing adaptability

Core Design Contradiction:

Device complexityVSAdaptability or versatility

Solution Approach 1:

The system dynamically selects which detection unit(s) to use based on the type of image processing to be performed. For processing requiring high accuracy (e.g., face recognition), only the first detection unit is used. For processing requiring comprehensive coverage (e.g., object trimming), both detection units operate and their results are combined. This dynamic adaptation resolves the contradiction by allowing the system to adjust its detection strategy according to processing needs.

Inventive Principle:

Principle #15Dynamics

Data Source

PatentUS9501688B2Apparatus, processing method and storage medium storing program

Publication Date: 2016.11.22 CANON KK

AI summary

An apparatus includes a detection unit configured to detect an object from image data, and a processing unit configured to perform processing for the object detected by the detection unit. The detection unit includes a first detection unit and a second detection unit having a detection tolerance wider than that of the first detection unit, and the object is detected by selecting the first detection unit or the second detection unit in accordance with the contents of the processing to be performed by the processing unit.