Red-Eye Correction Circuit with Status Suppression
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional image capturing devices, including digital cameras, face challenges in accurately and completely eliminating the red-eye effect during flash photography, as existing automatic correction methods often fail to detect and correct red-eyes in all instances, leading to erroneous corrections and unnatural image alterations.
Innovation Solution
An image-processing apparatus and method that includes a determination circuit to assess if an image has already undergone red-eye correction, and a correction suppression circuit to prevent further processing if the image has been corrected, ensuring accurate and targeted red-eye correction by allowing manual intervention when necessary.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If automatic red-eye correction is applied to all detected red regions, then more red-eyes are corrected, but erroneous corrections increase (e.g., correcting lips or already corrected regions)
Solution Approach 1:
The patent applies preliminary action by marking red-eye regions before correction and recording correction status in advance. The system detects red-eye regions, marks them with specific color information, and records which regions have been corrected. This preliminary marking and status recording prevents subsequent erroneous corrections of already corrected regions or non-red-eye regions that may be detected in later processing stages.
Solution Approach 2:
The patent implements feedback by using the correction status information and marked region data to control subsequent correction operations. The system continuously monitors which regions have been corrected and uses this feedback to suppress further correction in those areas, while allowing correction in uncorrected red-eye regions. This feedback mechanism ensures accurate and reliable correction by preventing erroneous re-correction.
2Manufacturing precision
If multiple correction passes are performed on the same image, then correction completeness improves, but re-correction errors increase
Solution Approach 1:
The patent records the correction status of each red-eye region in advance and maintains this information throughout multiple processing passes. By preliminarily marking corrected regions and storing this status, the system can distinguish between corrected and uncorrected regions in subsequent passes, allowing complete correction of missed red-eyes while preventing harmful re-correction of already corrected areas.
Solution Approach 2:
The system uses feedback from correction status records to control each correction pass. The correction status information feeds back into the detection and correction process, enabling the system to adjust its behavior in subsequent passes by suppressing correction in already corrected regions while maintaining correction capability for uncorrected regions, thus achieving completeness without re-correction errors.
3Speed
If red-eye correction is applied without checking prior correction status, then processing speed is maintained, but correction accuracy deteriorates
Solution Approach 1:
The patent performs preliminary marking of red-eye regions and preliminary recording of correction status before actual correction processing. This preliminary action organizes the data structure and correction status information in advance, enabling fast status checks during subsequent processing without significant speed penalty. The pre-organized data allows rapid determination of whether correction is needed, maintaining processing speed while improving accuracy.
Data Source
AI summary
This invention can more accurately correct a region in the screen, which suffers a red-eye effect. An image processing apparatus including a processing circuit which executes correction processing for correcting a region in a target image, which satisfies a condition that defines a correction target, comprises a determination circuit which determines whether the target image has already undergone the correction processing, and a correction suppression circuit which suppresses execution of the correction processing for the target image when said determination circuit determines that the target image has already undergone the correction processing.