Information Processing Device for 3D Line-of-Sight Estimation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Monitoring camera systems face challenges in accurately discriminating the state and information of individuals, especially when they are far from the imaging device or affected by light and shadows, leading to decreased recognition rates.
Innovation Solution
An information processing device and method that includes a detection system with a display device, imaging device, and processing unit to estimate the distribution of target information such as state and attributes by analyzing captured images, using units like acquisition, detection, discrimination, reliability calculation, estimation, and effect calculation to improve accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Area of stationary object
If a fixed imaging device is used to monitor persons in a large area, then the coverage area is increased, but the recognition accuracy of persons far from the imaging device decreases
Solution Approach 1:
The patent transitions from two-dimensional image analysis to three-dimensional spatial reasoning by estimating person positions in 3D space and calculating line-of-sight relationships. This dimensional enhancement allows the system to compensate for distance-related accuracy loss by incorporating spatial depth information and geometric relationships between the imaging device, persons, and advertisements.
Solution Approach 2:
The patent introduces an advertisement position database and line-of-sight calculation mechanism as intermediaries. These intermediaries bridge the gap between the imaging device and distant persons by providing reference points (advertisement positions) and computational methods (line-of-sight calculations) to determine whether distant persons can visually access advertisements, thereby maintaining recognition accuracy across large coverage areas.
2Quantity of substance
If the imaging device is positioned to cover a large area, then more persons can be monitored, but persons affected by light and shadow become harder to recognize
Solution Approach 1:
The patent replaces reliance on image quality (which degrades under light and shadow conditions) with computational geometry and line-of-sight analysis. Instead of depending on the mechanical/optical system to provide clear images of all persons, the system uses mathematical calculations based on position data to determine advertisement visibility, thereby maintaining reliable discrimination results regardless of lighting conditions.
Solution Approach 2:
The patent introduces advertisement position information and line-of-sight calculation as intermediaries that bypass the limitations of image-based recognition. By using these computational intermediaries, the system can determine whether persons are looking at advertisements based on geometric relationships rather than image quality, thus maintaining reliable discrimination even when light and shadow affect image clarity.
3Loss of information
If image processing is performed on all detected persons, then comprehensive data is obtained, but processing time and computational load increase
Solution Approach 1:
The patent applies local quality by performing detailed line-of-sight and visibility calculations only for persons who meet certain criteria (e.g., detected in the monitoring area, potential advertisement viewers) rather than uniformly processing all detected persons. This selective approach maintains information completeness for relevant targets while reducing unnecessary computational overhead for persons who cannot possibly be viewing advertisements.
Solution Approach 2:
The patent segments the processing workflow into distinct stages: first detecting persons in the monitoring area, then filtering based on position and visibility criteria, and finally performing detailed line-of-sight calculations only for candidate persons. This segmentation of the processing pipeline allows comprehensive data collection while minimizing processing time by avoiding redundant calculations on all detected persons.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
According to an arrangement, an information processing device includes an acquisition unit (42), a detection unit (44), a discrimination unit (46), and an estimation unit (50). The acquisition unit (42) is configured to acquire a captured image. The detection unit (44) is configured to detect a plurality of targets included in the captured image. The discrimination unit (46) is configured to calculate target information representing at least one of a state or an attribute and reliability of the target information for each of the plurality of detected targets on the basis of the captured image. The estimation unit (50) is configured to estimate a distribution of the target information of the plurality of targets on the basis of a distribution of the target information of targets for which the reliability is higher than a set value.