Iris Region Completion for Natural Facial Expressions
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing image processing technologies struggle to create natural-looking facial expressions due to incomplete iris detection, particularly when eyelids cover parts of the human eye, leading to unnatural eye movements and expressions.
Innovation Solution
An image processing device that detects the iris area in an image and fills in missing parts to create a substantially circular shape, using methods such as annular ring filters or color space conversions, to enhance the natural representation of eye movements and expressions.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Manufacturing precision
If conventional face image processing is used without iris completion, then the processing is simpler and faster, but the facial expression representation becomes unnatural
Solution Approach 1:
The patent applies preliminary action by completing the iris area before generating the final face image. The controller detects the iris area, identifies missing parts, and fills them in using reference information from the same eye or the other eye. This preprocessing step ensures that when the face image is generated, the eyes already have natural, complete appearances, leading to more realistic facial expressions without requiring complex post-processing adjustments.
2Measurement precision
If the iris area is detected and completed to be circular, then the eye representation becomes more natural, but the detection and processing becomes more complex
Solution Approach 1:
The patent applies parameter changes by transforming the iris area data into a standardized circular form. The controller detects the iris area, determines its parameters (center position, radius), and reconstructs it as a perfect circle using mathematical transformations. This parameter-based approach simplifies the detection process by focusing on key geometric parameters rather than complex pixel-level analysis, improving both accuracy and computational efficiency.
Solution Approach 2:
The patent applies copying by using reference information from the same eye or the other eye to complete missing iris parts. When part of the iris is obscured or incomplete in the detected image, the controller copies corresponding iris patterns and textures from visible portions of the same eye or from the other eye (which typically has similar characteristics). This copying approach accurately reconstructs the iris appearance without requiring complex de novo generation, enhancing detection precision while managing complexity.
Data Source
AI summary
The purpose of the present invention is to make a facial expression when the eyes are moved more natural. An image processing device (100) is provided with the following: an image acquisition unit (5a) for acquiring an image that includes the eyes; an iris detection unit (5c) for detecting from the acquired image the iris region of the eyes; and an eye image generation unit (5f) for supplementing a deficient portion of the iris region so that the shape of the detected iris region becomes substantially circular.