Gesture Recognition Text Input Trace Segmentation

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Conventional text recognition technologies face challenges in accurately recognizing text written in the air due to the consecutive nature of the writing trace, resulting in low precision.

Innovation Solution

A method and device utilizing hand gesture recognition that collects image information, determines the target object and writing plane, extracts the movement trace as a writing trace, and performs character recognition to improve recognition precision.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If conventional text recognition technology is used to recognize text written in the air, then the system can process consecutive writing traces, but the recognition precision is low

Engineering Contradiction:

Improverecognition precisionVSAvoidsystem complexity

Core Design Contradiction:

Measurement precisionVSDevice complexity

Solution Approach 1:

The patent segments the consecutive writing trace into discrete stroke units by identifying stroke boundaries through direction vector analysis and curvature detection. This segmentation transforms the continuous trace into separable character components, enabling accurate recognition while maintaining system feasibility.

Inventive Principle:

Principle #1Segmentation

Solution Approach 2:

The patent introduces intermediate processing modules including trace smoothing, direction vector calculation, and curvature detection between the raw consecutive trace and the final recognition system. These intermediary steps refine the trace data and extract meaningful features, improving recognition precision without requiring complete system redesign.

Inventive Principle:

Principle #24Intermediary (Mediator)

2Ease of manufacture

If the writing trace is treated as consecutive data, then the input method is simple to implement, but the stroke distinction and character recognition accuracy deteriorate

Engineering Contradiction:

Improveimplementation simplicityVSAvoidstroke distinction accuracy

Core Design Contradiction:

Ease of manufactureVSMeasurement precision

Solution Approach 1:

The patent performs preliminary processing on the writing trace before recognition, including smoothing the trace data, calculating direction vectors, and detecting curvature changes. These preliminary actions prepare the trace data for accurate stroke segmentation, maintaining implementation simplicity while improving stroke distinction accuracy.

Inventive Principle:

Principle #10Preliminary action

Solution Approach 2:

The patent transforms the consecutive trace data by changing parameters such as direction vectors, curvature values, and stroke boundary conditions. These parameter changes enable the system to distinguish between different strokes within the consecutive trace, improving recognition accuracy without complicating the overall implementation.

Inventive Principle:

Principle #35Parameter changes

Data Source

PatentUS10725552B2Text input method and device based on gesture recognition, and storage medium

Publication Date: 2020.07.28 SHENZHEN PRTEK

AI summary

A text input method based on gesture recognition, comprising: collecting image information containing a target object; determining the target object according to the image information, and acquiring a writing surface and a movement trace of the target object; extracting the movement trace of the target object in the writing surface as a writing trace and outputting same; and performing character recognition according to the writing trace, and outputting a character recognition result.