Particular-Region Detection for Red-Eye Correction

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Conventional image processing methods for detecting red eyes and golden eyes in images are time-consuming, particularly due to the need for accurate face detection in various orientations and sizes, which limits productivity in digital photoprinters.

Innovation Solution

A method and apparatus that detect particular regions like red eyes or golden eyes by changing detection conditions between main and non-main portions of an image, storing face information, and using it for subsequent face detection, thereby reducing unnecessary face detection and enhancing processing speed.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If conventional image processing methods are used to detect red eyes and golden eyes by performing face detection in various orientations and sizes, then detection accuracy is improved, but processing time increases significantly

Engineering Contradiction:

Improvedetection accuracyVSAvoidprocessing time

Core Design Contradiction:

Measurement precisionVSLoss of time

Solution Approach 1:

The image is divided into main portion and non-main portion regions. Face detection is performed with different conditions in each region: comprehensive detection in main portion and simplified detection in non-main portion, thereby reducing overall processing time while maintaining accuracy where needed

Inventive Principle:

Principle #1Segmentation

Solution Approach 2:

Different detection conditions are applied to different regions of the image. The main portion region receives full-face-detection treatment while the non-main portion region receives simplified treatment, optimizing the balance between accuracy and speed for each area

Inventive Principle:

Principle #3Local quality

2Reliability

If face detection is performed in all regions of the image to ensure accurate red eye and golden eye detection, then detection reliability is improved, but productivity decreases

Engineering Contradiction:

Improvedetection reliabilityVSAvoidprinter productivity

Core Design Contradiction:

ReliabilityVSProductivity

Solution Approach 1:

The image processing task is segmented into two regions with different processing intensities. The main portion region undergoes reliable face detection while the non-main portion uses simplified detection, ensuring productivity is maintained without sacrificing reliability in critical areas

Inventive Principle:

Principle #1Segmentation

Solution Approach 2:

Full face detection is performed only in the main portion region where it is most needed, while partial or simplified detection is applied in the non-main portion region. This partial action approach maintains sufficient reliability while significantly improving productivity

Inventive Principle:

Principle #16Partial or excessive action

Data Source

PatentUS7620242B2Particular-region detection method and apparatus, and program therefor

Publication Date: 2009.11.17 FUJIFILM CORP

AI summary

The particular region detecting method and apparatus detect particular region candidates in an image, then perform face detection in a region including one of the detected particular region candidates and specify as a particular region of a detection target the one candidate included in the region where a face can be detected. In the method and apparatus, detection conditions is changed for at least one of the candidate detection and the face detection between main and non-main portion regions of the image, face information regarding the face detected in the region is stored and the face detection corresponding to a particular region candidate to be next subjected to the face detection is performed using the thus stored face information. The program is caused a computer to execute the method.