Viewing Area Estimation Using Face Direction and Stationary Periods
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing viewing area estimation devices struggle to accurately estimate the viewing area of a monitored subject when image capturing conditions are affected by light sources or accessories like eyeglasses, making it difficult to detect the eyeball position.
Innovation Solution
A viewing area estimation device that uses a captured image acquisition portion to acquire multiple images of a subject's face, a face direction detection portion to identify the face direction and stationary periods, and a direction estimation portion to estimate the viewing area based on these parameters, allowing for reliable estimation even when the eyeball position is not recognizable.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If eyeball position detection is used to estimate viewing area, then measurement precision is improved, but reliability deteriorates when image capturing conditions are affected by light sources or eyeglasses
Solution Approach 1:
The patent introduces face direction as an intermediary parameter to estimate viewing area when direct eyeball position detection fails. Instead of directly detecting eyeball positions (which fails under certain lighting conditions), the system detects face direction from multiple captured images and uses this as a substitute indicator to infer the monitored subject's viewing area, thereby maintaining reliability under various image capturing conditions
Solution Approach 2:
The patent changes the detection parameter from eyeball position to face direction. By detecting face direction (orientation of the face) instead of eyeball position, the system overcomes the limitation of eyeball detection under poor lighting conditions or when eyeglasses are worn, while still enabling viewing area estimation through the relationship between face direction and viewing area
2Measurement precision
If eyeball position detection is used, then viewing area estimation accuracy is improved, but ease of operation deteriorates due to difficulty in detecting eyeball positions under various conditions
Solution Approach 1:
The system uses face direction as an intermediary that is easier to detect than eyeball position. Face direction can be reliably determined from facial features in captured images even when eyeball detection fails, making the overall detection process more robust and easier to perform across various conditions
3Reliability
If face direction detection is used instead of eyeball position detection, then reliability is improved, but measurement precision may deteriorate
Solution Approach 1:
The system uses face direction as a partial indicator of viewing area. While face direction alone may not provide complete viewing area information, it provides sufficient reliability for estimation purposes, especially when combined with the understanding that face direction correlates with general viewing direction even if not perfectly precise
Data Source
AI summary
A viewing area estimation device, which estimates a viewing area visually confirmed by a monitored subject, includes a captured image acquisition portion acquiring a plurality of chronologically captured images of a face of a monitored subject, a face direction detection portion detecting, based on the plurality of captured images, a face direction of the monitored subject and a stationary period representing a duration in which a face direction change of the monitored subject is in a stop state or in an almost stop state, and a direction estimation portion estimating the viewing area based on parameters including the face direction of the monitored subject and the stationary period of the face direction change.