Interactive Projector Sub-Pixel Pointing Detection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing interactive projectors face challenges in accurately detecting the tip position of a pointing element, which is crucial for user instructions, due to limitations in detection methods and accuracy.
Innovation Solution
An interactive projector system that uses a correlation value calculation section to generate a correlation value map by comparing template images with taken images, allowing for the detection of the pointing element's tip position with accuracy finer than one pixel, and includes features like resolution-converted images and local maximum value detection to enhance detection precision.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If template matching is performed using object templates to detect the pointing element, then the detection process can be implemented, but the detection accuracy of the tip position cannot be improved beyond one pixel resolution
Solution Approach 1:
The patent segments the detection process into multiple stages: first performing template matching to identify the pointing element, then calculating correlation values for each pixel, and finally determining the tip position through centroid calculation. This segmentation allows achieving sub-pixel accuracy without requiring a complete redesign of the detection system.
Solution Approach 2:
The patent transitions from discrete pixel-level detection to continuous sub-pixel level detection by introducing correlation value calculations. By computing correlation values and their centroids, the system effectively adds a dimensional precision beyond the discrete pixel grid, achieving accuracy corresponding to values smaller than one pixel.
2Measurement precision
If the tip position is detected by calculating correlation values between template images and taken images, then detection accuracy can be improved, but the calculation complexity and processing time increase
Solution Approach 1:
The patent performs preliminary template matching to identify the pointing element before conducting the more computationally intensive correlation value calculations. This preliminary action filters out non-relevant areas, allowing the detailed correlation analysis to be focused only on regions containing the pointing element, thereby improving processing efficiency.
Solution Approach 2:
The patent replaces traditional mechanical or discrete pixel-based detection methods with correlation-based mathematical operations. By using correlation value calculations and centroid computations, the system achieves continuous precision without relying on discrete pixel boundaries, effectively substituting a more sophisticated mathematical approach for simpler but less accurate methods.
3Adaptability or versatility
If multiple template images are used to detect different pointing elements, then detection versatility is improved, but the complexity of generating and managing correlation value maps increases
Solution Approach 1:
The patent employs multiple template images that can match different types of pointing elements (e.g., fingers, pens, other pointing devices). This universal approach allows the same detection system to handle various pointing element types by selecting or matching against appropriate templates, thereby achieving versatility without requiring separate detection systems for each element type.
Data Source
AI summary
An interactive projector includes an imaging section adapted to take an image of a projected screen, a correlation value calculation section adapted to calculate a correlation value between each of template images used to detect the pointing element and a taken image taken by the imaging section and including the pointing element to generate a first correlation value for each of the template images, a correlation value map generation section adapted to generate a correlation value map representing the correlation values between the plurality of template images and the taken image based on the plurality of first correlation values generated for the respective template images, and a pointing element detection section adapted to extract a high correlation value area in the correlation value map, and calculate a centroid of the correlation values in the high correlation value area to thereby detect a tip position of the pointing element.