Touch Sensor Contact Segmentation for Accidental Touch Classification
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing touch sensor devices struggle to accurately classify different types of contacts, particularly distinguishing between intentional and accidental touches, especially in larger touch regions and when users interact with keyboards, leading to a degraded user experience.
Innovation Solution
A neural network-based classification system that processes touch data from sensor electrodes to generate touch images, identify contact segments, and classify contacts as intentional or accidental, using techniques like data pre-processing and augmentation to enhance accuracy in edge and multi-contact scenarios.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Area of stationary object
If the touch region size is increased, then the active usable area is improved, but the accuracy of contact classification deteriorates due to increased accidental palm contacts
Solution Approach 1:
The patent segments the touch contact into multiple contact segments by identifying local maxima in the touch image and creating segmented touch images. Each segment is then independently classified using the neural network. This segmentation allows the system to distinguish between intentional contacts (typically single, localized) and accidental palm contacts (typically larger, distributed across multiple segments), thereby maintaining classification accuracy even in larger touch regions.
2Device complexity
If traditional touch classification methods are used, then the device complexity is low, but the ability to distinguish intentional from accidental contacts deteriorates
Solution Approach 1:
The patent replaces traditional rule-based or threshold-based touch classification methods with a neural network-based machine learning system. The neural network is trained on labeled touch data to recognize patterns distinguishing intentional from accidental contacts. This substitution significantly improves classification reliability by learning complex patterns from data, while the use of pre-trained models and efficient inference techniques keeps the operational complexity manageable.
3Measurement precision
If more touch features are analyzed to improve classification accuracy, then the measurement precision is improved, but the processing time and computational resources increase
Solution Approach 1:
The patent performs preliminary processing of the touch image before neural network classification, including identifying local maxima, creating segmented touch images, and extracting relevant features. By pre-processing and segmenting the touch data, the system reduces the complexity of the classification task and enables faster neural network inference. The segmentation approach also allows parallel processing of multiple contact segments, further reducing overall processing time.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
An input device for classification of an input object is provided. The input device comprises a touch sensor comprising a plurality of sensor electrodes configured to obtain touch data; and a processing system. The processing system is configured to receive touch data from resulting signals from the plurality of sensor electrodes; generate a touch image based on the touch data; generate one or more contact images based on the touch image, each contact image comprising one or more first pixels from the touch image and one or more second pixels with predefined values; classify, using a neural network, a respective contact in each of the one or more contact images and generate corresponding classification results; and identify, based on the classification results, one or more classified contacts in the touch image.