Motion Recognition Apparatus Cyclicity Loss Detection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing motion recognition systems face challenges in efficiently recognizing short-duration 'event motions' such as 'stand up' or 'sit' without incurring significant calculation overhead, particularly on devices with limited resources like mobile phones, due to the need for frequent data extraction and processing.
Innovation Solution
A motion recognizing apparatus and method that detects cyclicity loss in sensor data to set optimal data sections for recognition, reducing the number of processing cycles and calculation required for 'event motions' by predicting peak intervals and matching them with actual peaks.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If time window data extraction and recognition processing are performed frequently to detect short-duration event motions, then motion recognition accuracy is improved, but calculation amount and processing load increase significantly
Solution Approach 1:
The system performs preliminary analysis by detecting cyclicity loss in sensor data to identify potential event motion occurrence times before performing full recognition processing. This preliminary detection step filters out time windows that do not contain event motions, so that subsequent detailed recognition processing is only performed on selected time windows where event motions are likely to occur, significantly reducing the overall calculation amount while maintaining recognition accuracy
Solution Approach 2:
The recognition process is divided into two stages: first, cyclicity loss detection is performed on continuous sensor data to identify candidate time windows; second, full recognition processing is performed only on the selected candidate time windows. This segmentation allows the system to maintain high recognition accuracy for short-duration event motions while avoiding the computational overhead of performing full recognition processing on all possible time windows
2Reliability
If the time window duration is extended to ensure complete capture of short-duration event motions, then detection reliability is improved, but processing time and energy consumption increase
Solution Approach 1:
Cyclicity loss detection is performed as a preliminary step to precisely identify the start times of event motions before extracting time window data. By detecting the loss of cyclicity pattern in sensor data, the system can accurately determine when an event motion occurs, allowing for optimized time window extraction that captures the complete event motion without requiring excessively long time windows, thus reducing processing time while maintaining detection reliability
Data Source
AI summary
Provided are a motion recognition apparatus, a motion recognition system and a motion recognition method that enable‘event motions’ to be recognized with a small number of calculations. The motion recognition system, which recognizes user motions by using sensor data, is configured to be provided with: a cyclical loss detection means for detecting cyclical losses of sensor data when a user is moving; and a recognition processing means for setting data intervals to be used for recognizing motions in accordance with the cyclical losses of sensor data that were detected, and for recognizing user motions on the basis of sensor data for the data intervals that have been set.