SOM State Estimation Handling Discontinuous Image Regions
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional Self-organizing Map (SOM) techniques face difficulties in performing classification processing and pattern determination when the generated map includes discontinuous image regions, making it challenging to accurately classify input waveform data.
Innovation Solution
A state determination apparatus comprising feature vector obtaining circuitry, mapping conversion circuitry, matching processing circuitry, and state determination circuitry is developed to obtain feature vectors, map them into a different dimensional space, calculate adaptability data, and determine the state indicated by the measured data, even in the presence of discontinuous image regions.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If conventional SOM technique is used to generate waveform map, then continuous image regions can be formed, but discontinuous image regions (split regions) cannot be handled for classification processing
Solution Approach 1:
The patent divides the waveform map into multiple candidate regions that may be discontinuous or split. Instead of requiring a single continuous region, the system segments the map into multiple potential classification regions and evaluates each independently through template matching, allowing discontinuous regions to be properly handled and classified.
Solution Approach 2:
The patent introduces template data as an intermediary between the waveform map and classification results. By comparing candidate regions against predefined templates and calculating adaptability data, the system can accurately classify discontinuous regions without requiring them to be continuous, thus resolving the contradiction between handling discontinuous regions and maintaining classification accuracy.
2Measurement precision
If template matching with adaptability data calculation is performed, then accurate state determination is achieved, but additional processing steps are required
Solution Approach 1:
The patent performs preliminary actions by pre-defining template data representing different waveform patterns before actual classification. These templates are prepared in advance and stored, allowing the system to quickly compare candidate regions against known patterns without complex real-time analysis, thus achieving accurate state determination with manageable processing complexity.
Solution Approach 2:
The patent changes the parameter of comparison by using adaptability data that quantifies the match between candidate regions and templates. By transforming the classification problem into a parameter-based adaptability calculation, the system achieves precise state determination through straightforward numerical comparison rather than complex pattern recognition, balancing accuracy and processing complexity.
Data Source
AI summary
Provided is a state determination apparatus that appropriately performs pattern classification processing and/or pattern determination processing even when a map generated by the SOM technique includes discontinuous image regions (e.g., split image regions). In the state determination apparatus, the matching processing unit obtains adaptability data indicating a correlation degree between template data indicating a predetermined state and the SOM output data. The state determination unit determines a state of an input data. This allows for appropriately performing pattern classification processing and/or pattern determination processing even when a map generated by the SOM technique includes discontinuous image regions (e.g., split image regions).