Touch Input Classification for Stylus and Finger Differentiation

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Conventional devices and touch surfaces fail to differentiate between intended and unintended touch inputs simultaneously with an input tool, leading to inefficient processing and potential accidental actions.

Innovation Solution

A multi-stage classification process that evaluates various input factors such as inter-arrival time, inter-arrival distance, number of inputs, direction, and velocity to distinguish between tool inputs and user touch inputs, allowing for confident classification of intended and unintended touch inputs.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Productivity

If the device processes all touch inputs without differentiation, then no classification is performed, but processing efficiency decreases and accidental actions increase

Engineering Contradiction:
Improveprocessing efficiencyVSAvoidaccuracy of input recognition
Core Design Contradiction:
ProductivityVSReliability

Solution Approach 1:

The patent segments touch inputs into multiple categories (intended tool inputs, unintended palm contacts, unintended finger contacts) based on spatial relationships and temporal patterns. By dividing the input processing into distinct classification categories, the system can efficiently route different input types to appropriate handling procedures, improving both processing efficiency and accuracy simultaneously

Inventive Principle:
Principle #1Segmentation

2Device complexity

If the device implements a simple touch input handling system, then device complexity is low, but the ability to differentiate between intended and unintended inputs is insufficient

Engineering Contradiction:
Improvecomplexity of input handling systemVSAvoidprecision of input classification
Core Design Contradiction:
Device complexityVSMeasurement precision

Solution Approach 1:

The patent utilizes multiple parameters including spatial coordinates, temporal sequences, velocity vectors, and pressure data to classify touch inputs. By changing from simple binary classification to multi-parameter analysis, the system achieves high classification precision while managing complexity through systematic parameter evaluation

Inventive Principle:
Principle #35Parameter changes

3Measurement precision

If the device uses multiple classification factors and multi-stage classification, then input classification accuracy improves, but processing time and computational resources increase

Engineering Contradiction:
Improveclassification accuracyVSAvoidprocessing time
Core Design Contradiction:
Measurement precisionVSLoss of time

Solution Approach 1:

The patent performs preliminary classification using readily available spatial and temporal data before more complex analysis. By establishing initial classifications based on basic parameters and only proceeding to deeper analysis when necessary, the system minimizes processing time while maintaining high classification accuracy for the majority of inputs

Inventive Principle:
Principle #10Preliminary action

Data Source

PatentEP3191922B1Classification of touch input as being unintended or intended
Publication Date: 2019.04.24 MICROSOFT TECHNOLOGY LICENSING LLC
  • EP3191922B1 patent drawingFigure 1
  • EP3191922B1 patent drawingFigure 2
  • EP3191922B1 patent drawingFigure 3

AI summary

The techniques described herein implement a classification process to evaluate information associated with a tool input (e.g., from an input tool such as a pen or a stylus) and a user touch input (e.g., from a finger or a palm of a hand) to determine whether the user touch input is an intended or an unintended touch input. The information evaluated may be associated with an arrival of the tool input relative to an arrival of the user touch input. The information evaluated may also be associated with a movement of the tool input relative to a movement of the user touch input. In various implementations, the techniques may calculate an evaluation score and compare the evaluation score to a confidence classification threshold. If a confident classification cannot be achieved, the techniques further the classification process as more information associated with the inputs is received.