Template-Matched Eye Detection for Blinking 3D Displays
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing eye position detection methods struggle to accurately determine pupil positions, especially when images do not include pupils, such as during blinking, leading to computation inefficiencies and interruptions in detection.
Innovation Solution
A detection device employs template matching using first and second template images to detect eye positions, where the first image focuses on eyes and the second on the face, allowing for accurate detection even when pupils are not visible, and outputs coordinate information for 3D projection systems.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If traditional pupil detection methods are used to detect eye positions, then detection accuracy is maintained when pupils are visible, but detection fails or requires complex processing when pupils are not visible (e.g., during blinking)
Solution Approach 1:
The detection process is segmented into two distinct template matching processes: first template matching for detecting eye positions when pupils are visible, and second template matching for detecting eye positions when pupils are not visible. This segmentation allows each process to be optimized for its specific condition, improving overall reliability without requiring a single complex detection system.
Solution Approach 2:
The invention changes the detection parameters by switching between different template images based on detection conditions. When pupils are not visible, the system changes from pupil-centered detection to face-region-based detection, allowing continuous eye position tracking even during blinking without increasing device complexity.
2Measurement precision
If traditional pupil detection methods are used, then accurate eye positions can be detected when pupils are visible, but computational efficiency decreases and detection interruptions occur when pupils are not visible
Solution Approach 1:
The system performs preliminary template matching using a first template image focused on eye regions to quickly determine if pupils are visible. Based on this preliminary detection, it pre-prepares the appropriate second template image for face region matching, ensuring continuous accurate detection without computational interruptions during blinking.
Solution Approach 2:
The detection system dynamically switches between different template matching approaches based on real-time detection conditions. When pupils are visible, it uses the faster first template matching; when pupils are not visible, it transitions to the second template matching method, maintaining both accuracy and efficiency across varying conditions.
3Device complexity
If a single template matching process is used for eye detection, then the detection process is simple, but accuracy decreases when pupils are not visible in the image
Solution Approach 1:
The detection system achieves multi-functionality by implementing two template matching processes that can handle different detection scenarios. The first process handles normal eye detection, while the second process handles detection during blinking or when pupils are not visible. This universal approach maintains simplicity while improving accuracy across various conditions.
Data Source
AI summary
A detection device includes an input device and a detector. The input device in the detection device receives input of image information output from a camera. The detector performs a detection process to detect positions of eyes of a user. The detector performs, as the detection process, a first process to detect first positions of the eyes based on the image information by template matching, and a second process to detect a position of a face based on the image information by template matching and detect second positions of the eyes based on the detected position of the face.