Motion Analysis Apparatus Contribution Degree Determination
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional motion analysis apparatuses require significant time and effort to set a suitable contribution degree for analysis conditions, especially when dealing with unsteady motions, necessitating extensive experimentation and data accumulation.
Innovation Solution
A motion analysis apparatus and method that extract unsteady motions using a moving image, featuring a motion data input section, motion feature extraction, principal component analysis, distance calculation, and contribution degree determination sections to easily set the contribution degree by calculating distances from learning and non-learning data.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If the contribution degree is set high to prevent abnormal motion determination leakage, then the detection reliability is improved, but the risk of erroneous determination of standard motion increases
Solution Approach 1:
The patent implements feedback by calculating the distance between learning data and the subspace, then using this distance information to determine the suitability of the contribution degree. The system continuously adjusts the contribution degree based on feedback from distance calculations, ensuring optimal balance between detecting abnormal motions and avoiding false positives for standard motions.
Solution Approach 2:
The patent changes the contribution degree parameter dynamically based on analysis conditions. By adjusting this parameter according to the calculated distances and suitability evaluation, the system optimizes the balance between sensitivity to abnormal motions and tolerance for normal motion variations.
2Measurement precision
If extensive experimentation and data accumulation are performed to set the optimal contribution degree, then the detection accuracy is improved, but the time and effort required increases significantly
Solution Approach 1:
The patent performs preliminary action by automatically determining the suitable contribution degree using distance calculations from learning data before actual motion analysis begins. This preliminary determination eliminates the need for extensive experimentation and data accumulation during deployment, saving significant time and effort.
Solution Approach 2:
The system performs self-service by automatically evaluating the suitability of contribution degrees through distance calculations and determining the optimal value without requiring external experimentation or manual tuning. The apparatus independently configures itself based on the learning data provided.
Data Source
AI summary
A motion analysis apparatus is provided that enables a contribution degree that suits analysis conditions to be set easily. A motion analysis apparatus (300) is provided with a motion data input section (310) that receives learning data as input, a motion feature extraction section (320) that extracts a motion feature amount from learning data, a principal component analysis section (330) that performs principal component analysis using a motion feature amount on part of the learning data, and learns a subspace, a learning data distance calculation section (340) that calculates a distance between a learning data motion feature amount and a subspace, and a contribution degree determination section (350) that determines the suitability of a contribution degree used in principal component analysis, from a distance calculated from learning data that is used in subspace learning and a distance calculated from learning data that is not used in subspace learning.