Multi-Input Target Prediction Engine for Gesture Error Correction

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Current user interfaces for electronic devices, such as mobile phones and tablets, lack error correction in user selection events for multi-input targets, leading to inaccurate input due to factors like user distraction, interface layout, and gesture accuracy, resulting in a frustrating user experience.

Innovation Solution

A system and method that model user selection events using probability distributions, such as sigmoid and Gaussian functions, to predict the intended input option, allowing for error correction and more accurate predictions by analyzing the time and location of user gestures, like drag-and-lift and multi-tap interactions.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Reliability

If deterministic selection is used for multi-input targets, then the system is simple and fast, but input accuracy deteriorates due to user errors like dragging in the wrong direction or lifting too soon

Engineering Contradiction:
Improveinput accuracyVSAvoidsystem complexity
Core Design Contradiction:
ReliabilityVSDevice complexity

Solution Approach 1:

The system implements feedback by continuously monitoring the user's drag gesture progress and providing predictive suggestions for the intended character. The prediction engine analyzes the current gesture state and compares it with expected gesture patterns, then provides feedback in the form of suggested corrections or confirmations to the user, allowing them to verify or adjust their intended input before final submission.

Inventive Principle:
Principle #23Feedback

Solution Approach 2:

The system performs preliminary action by predicting the user's intended character selection during the drag gesture, before the user actually completes the selection. The prediction engine proactively generates likely character candidates based on the partial gesture information and prepares suggestions in advance, so that when the user completes the gesture or makes an error, the system can quickly offer corrections without waiting for the final action.

Inventive Principle:
Principle #10Preliminary action

2Reliability

If error correction modeling is added to user selection events, then input accuracy improves, but processing time increases due to probability calculations

Engineering Contradiction:
Improveinput accuracyVSAvoidprocessing time
Core Design Contradiction:
ReliabilityVSLoss of time

Solution Approach 1:

The system applies partial action by performing probability calculations and error correction modeling only for gestures that show signs of potential errors or ambiguity. For clear, unambiguous gestures that confidently indicate the intended character, the system skips the computationally intensive probability calculations and directly processes the input, thus avoiding unnecessary processing time while still correcting errors when they occur.

Inventive Principle:
Principle #16Partial or excessive action

Solution Approach 2:

The system changes parameters dynamically by adjusting the threshold for triggering error correction based on gesture confidence levels. When the gesture parameters (duration, distance, speed, path) indicate high confidence in the intended character, the system reduces or eliminates probability calculations. When parameters suggest uncertainty or potential error, the system increases the level of probabilistic analysis, thus optimizing processing time based on the actual need for correction.

Inventive Principle:
Principle #35Parameter changes

3Reliability

If the user interface requires precise gesture execution, then input accuracy improves, but ease of operation deteriorates due to user distraction or instability

Engineering Contradiction:
Improveinput accuracyVSAvoidgesture simplicity
Core Design Contradiction:
ReliabilityVSEase of operation

Solution Approach 1:

The system provides continuous feedback during the drag gesture by displaying predicted character suggestions as the user drags. This feedback allows users to see what character the system thinks they intend to select, giving them confidence that their gesture is being interpreted correctly. If the prediction is wrong, users can adjust their gesture or select the suggested character, making the interface more forgiving of imprecise gestures while maintaining accuracy.

Inventive Principle:
Principle #23Feedback

Solution Approach 2:

The prediction engine acts as an intermediary between the user's gesture and the final character selection. Instead of directly mapping the gesture to a character (which requires high precision), the intermediary prediction layer interprets the gesture, considers multiple possible intentions, and provides a best-guess selection that the user can then confirm or correct. This intermediary step buffers against gesture imprecision and user distraction.

Inventive Principle:
Principle #24Intermediary (Mediator)

Data Source

PatentEP3114546B1System and method for inputting one or more inputs associated with a multi-input target
Publication Date: 2024.01.03 MICROSOFT TECHNOLOGY LICENSING LLC
  • EP3114546B1 patent drawingFigure 1a~1c
  • EP3114546B1 patent drawingFigure 2a~2c
  • EP3114546B1 patent drawingFigure 3a~3c

AI summary

The present provides a system comprising a user interface and a prediction engine. The user interface comprises a multi-input target associated with multiple input options available for input by user selection. In a first system, the user interface is configured to select the multi-input target upon receipt of a first user selection event; and select an input of the multiple input options upon receipt of a second user selection event, wherein the second user selection event is spatially separated from the first user selection event. The prediction engine is configured to determine, based on the second user selection event, the probability that the user had intended to select an input of the multiple inputs other than the selected input. In a second system, the user interface is configured to identify a first input of the multiple input options upon receipt of a first user selection event directed at the multi-input target;and change the identified input from the first input to a second input of the multiple input options upon receipt of a second user selection event directed at the multi-input target. The prediction engine is configured to determine the time elapsed between the first and second user selection events; determine that the time elapsed is less than a predetermined threshold; and estimate, based on the elapsed time, the probability that the user had intended to select the second input and estimate the probability that the user had intended to input the first input and then select the first input again. The predetermined threshold is set such that if the elapsed time is greater than the predetermined threshold, the first input is input into the system with a probability of 1, and the first input is selected again. Corresponding methods are also provided.