Multi-touch Recognition Using Blur Filtering and Adaptive Background
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing multi-touch recognition systems face noise interference from infrared camera images, screen background noise, and wavelength-specific light issues, leading to reduced touch recognition accuracy and sensitivity.
Innovation Solution
A multi-touch recognition apparatus that employs blur filtering and adaptive background estimation to remove noise and background interference from images captured by cameras on a display panel, using a controller to calculate and update background images based on difference images and binary images, thereby enhancing touch recognition performance.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If infrared camera is used for touch recognition, then touch recognition can be achieved on display panel, but noise in image leads to touch recognition errors
Solution Approach 1:
The system performs preliminary background image acquisition and storage before actual touch recognition. By capturing a background image without touch input and storing it in advance, the system prepares reference data that will be used to subtract noise and interference from subsequent touch images, thereby improving touch recognition accuracy
Solution Approach 2:
The system extracts and removes the background image from the captured touch image through image subtraction. By separating the background component from the foreground touch elements, the system eliminates noise and interference patterns that would otherwise cause recognition errors, leaving only the relevant touch information
2Loss of information
If screen background image is included in recognition, then complete image information is available, but screen itself serves as noise reducing sensitivity
Solution Approach 1:
The system extracts the background image component from the composite image by performing image subtraction between the captured image and the stored background image. This separation removes the screen background and display elements that act as noise, isolating only the touch-related information for accurate recognition
Solution Approach 2:
The system segments the captured image into background components and foreground touch components through subtraction operations. By dividing the image information into separate elements, the system can process only the relevant touch portions while discarding the interfering background elements
3Adaptability or versatility
If light of specific wavelength enters infrared camera, then camera may function with broader light sources, but saturation occurs and image distortion prevents recognition
Solution Approach 1:
The system continuously updates the background image based on current image input and detected changes. By using feedback from the difference image and binary image to recalculate and update the background, the system adapts to changing lighting conditions and maintains reliable recognition even when wavelength-specific light is present
Solution Approach 2:
The background image is not static but dynamically updated based on current imaging conditions. The system adjusts the background reference continuously through weighted calculations involving the binary image, allowing the system to adapt to varying light sources and environmental conditions while maintaining recognition accuracy
Data Source
Figure 1
Figure 2
Figure 3
AI summary
Disclosed herein are a multi-touch recognition apparatus and a control method thereof. The multi-touch recognition apparatus executes blur filtering for noise removal when an image is input through a camera unit photographing a screen of a display panel supporting multi-touch, calculates and outputs a difference image obtained by removing the background image stored in the storage unit from the blur-filtered image, calculates a new background image using the difference image and a binary image of the difference image, and updates the background image stored in the storage unit using the calculated new background image, thereby effectively removing a background other than multi-touch and thus improving multi-touch recognition performance.