Motion Data Labelling via Reference Model Comparison

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Conventional methods for collecting and labeling motion data are cumbersome, time-consuming, and require excessive human intervention, leading to inaccuracies, especially in continuous motion data applications like aerobic exercises.

Innovation Solution

A device and method that automatically label motion data by comparing motion signals with a reference model and a motion script, using a processor to determine similarities and label data groups based on preset time messages and thresholds, reducing human intervention and improving accuracy.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Extent of automation

If conventional single collection and labelling method is used, then motion data can be collected and labelled, but the process requires excessive human intervention and is time-consuming

Engineering Contradiction:

Improveautomation of motion data labellingVSAvoidtime required for collection and labelling

Core Design Contradiction:

Extent of automationVSLoss of time

Solution Approach 1:

The system automatically labels motion data by comparing motion signals with reference models and motion scripts without requiring manual intervention. The processor autonomously determines similarity scores, compares them against thresholds, and generates labels, making the system self-sufficient in the labelling process.

Inventive Principle:

Principle #25Self-service

Solution Approach 2:

Reference models and motion scripts are pre-prepared and stored in the system before actual motion data collection. These reference materials contain predetermined motion patterns and time messages that enable automated comparison and labelling, eliminating the need for manual labelling during data collection.

Inventive Principle:

Principle #10Preliminary action

2Measurement precision

If conventional batch collection method is used, then multiple motion data can be collected at once, but the accuracy of labelling is not good enough

Engineering Contradiction:

Improveaccuracy of motion data labellingVSAvoidcomplexity of labelling process

Core Design Contradiction:

Measurement precisionVSDevice complexity

Solution Approach 1:

The system calculates similarity scores between motion signals and reference models, compares these scores against predetermined thresholds, and uses this feedback to determine whether to apply labels. This feedback mechanism ensures accurate labelling by objectively evaluating how well motion data matches reference patterns.

Inventive Principle:

Principle #23Feedback

Solution Approach 2:

The labelling process is divided into distinct steps: determining subsets of motion signals based on time message matching, calculating similarity scores for each signal, comparing scores against thresholds, and applying labels only to signals that meet the criteria. This segmented approach improves accuracy while maintaining manageable complexity.

Inventive Principle:

Principle #1Segmentation

3Productivity

If manual labelling method is used, then motion data can be labelled, but the process is cumbersome and requires excessive human intervention

Engineering Contradiction:

Improveefficiency of motion data labellingVSAvoidease of motion data collection and labelling

Core Design Contradiction:

ProductivityVSEase of operation

Solution Approach 1:

The manual mechanical process of human labelling is replaced with an automated electronic system. The processor electronically compares motion signals with reference models, calculates similarity scores, and automatically generates labels, substituting human manual operations with automated computational processes.

Inventive Principle:

Principle #28Mechanics substitution (Replace mechanical system)

Solution Approach 2:

The system uses reference models that are copies or representations of standard motion patterns. By comparing actual motion signals against these reference copies, the system can automatically determine labels without requiring human experts to manually analyze each motion data point.

Inventive Principle:

Principle #26Copying

Data Source

PatentUS11164094B2Device, method, and non-transitory computer readable storage medium for labelling motion data

Publication Date: 2021.11.02 INSTITUTE FOR INFORMATION INDUSTRY

AI summary

A device, method, and non-transitory computer readable storage medium for labelling motion data are provided. The device receives several motion signals, wherein each motion signal includes a motion time message and a motion data group. A motion script includes a plurality of preset motion messages, wherein each preset motion message includes a preset time message and a preset motion. The device performs the following steps for each preset time message: determining a first subset of the motion signals by comparing the motion time messages with the preset time message, calculating a similarity between the motion data group of each motion signal in the first subset and a reference model, determining a second subset of the first subset based on the first similarities, and labelling the motion data group of each motion signal included in the second subset as corresponding to the preset motion corresponding to the preset time message.