AR Navigation Line of Sight Detection via Facial Feature Segmentation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing interactive navigation systems struggle to accurately determine the line of sight of users when they are far away or when multiple users have different viewing angles, leading to incorrect display of virtual information on transparent displays.
Innovation Solution
An active interactive navigation system that includes a light-transmittable display device, a target object image capturing device, a user image capturing device, and a processing device, which recognizes users, detects their line of sight, and calculates the cross-point position for displaying virtual information by using facial feature matching and image cutting techniques to improve face recognition and prioritize service users.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Area of stationary object
If the user is far away from the display device, then the field of view coverage is improved, but the line of sight detection accuracy deteriorates
Solution Approach 1:
The patent divides the user image into multiple sub-images (first sub-image, second sub-image, etc.) to perform face recognition on segmented regions. This segmentation approach allows the system to maintain detection accuracy even when users are far away by focusing computational resources on relevant facial regions rather than processing the entire large-scale image.
Solution Approach 2:
The patent applies different processing strategies to different regions of the image. For far users, it performs image cutting to extract facial regions; for near users, it uses direct face recognition. This local quality adjustment optimizes detection accuracy for each specific viewing distance scenario.
2Adaptability or versatility
If multiple users are viewing the dynamic objects at the same time, then the system coverage is improved, but the line of sight determination accuracy deteriorates
Solution Approach 1:
The patent segments the image based on user positions and divides it into multiple sub-images, each corresponding to a specific user's viewing area. This segmentation enables the system to independently analyze and determine the line of sight for each user, maintaining accuracy even when multiple users are present simultaneously.
Solution Approach 2:
The patent applies user-specific processing to different regions of the image. Each user's facial features and line of sight are detected and analyzed independently in their respective image sub-regions, allowing the system to maintain high determination accuracy for each user while supporting multiple concurrent users.
3Measurement precision
If image cutting is performed to improve face recognition, then the recognition accuracy is improved, but the processing time increases
Solution Approach 1:
The patent performs image cutting and facial feature extraction as preliminary actions before the main face recognition process. By pre-processing the image to extract only the relevant facial regions, the system reduces the computational burden of the subsequent recognition step, thereby minimizing overall processing time while maintaining high accuracy.
Solution Approach 2:
The patent extracts only the necessary facial features and regions from the full image for recognition purposes. By taking out and processing only the relevant facial information rather than the entire image, the system achieves high recognition accuracy with reduced processing time.
Data Source
AI summary
An active interactive navigation system includes a display device, a target object image capturing device, a user image capturing device, and a processing device. The target object image capturing device captures a dynamic object image. The user image capturing device obtains a user image. The processing device recognizes and selects a service user from the user image and captures a facial feature of the service user. If the facial feature matches facial feature points, the processing device detects a line of sight of the service user and accordingly recognizes a target object watched by the service user, generates face position three-dimensional coordinates corresponding to the service user, position three-dimensional coordinates corresponding to the target object, and depth and width information, accordingly calculates a cross-point position where the line of sight passes through the display device, and display virtual information of the target object on the cross-point position of the display device.