Feature Space Closed Region Category Determination

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Conventional determination devices face challenges in accurately classifying objects in images based on feature values, as they often rely on complex probability density functions or representative data, which can be inefficient and less accurate, especially when dealing with complex data distributions.

Innovation Solution

A determination device and method that utilize a region information recording unit to store closed regions formed by nodes and line segments within a feature space, allowing for category decision based on the position of the determination target within these regions, using self-organizing feature maps to calculate distances and accurately represent data distributions with less representative data.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Adaptability or versatility

If conventional determination devices use complex probability density functions or representative data to classify objects, then they can handle various data distributions, but the computational complexity increases and determination time is extended

Engineering Contradiction:

Improveability to handle various data distributionsVSAvoidcomputational complexity

Core Design Contradiction:

Adaptability or versatilityVSDevice complexity

Solution Approach 1:

The feature space is segmented into multiple closed regions, each corresponding to a specific category. By dividing the continuous feature space into discrete regions with clear boundaries, the system avoids complex probability density calculations while maintaining the ability to handle various data distributions. Each closed region is defined by connecting nodes with line segments, creating a segmented representation that simplifies classification.

Inventive Principle:

Principle #1Segmentation

Solution Approach 2:

The patent transitions from traditional probability-based classification in feature space to a geometric dimension by introducing closed regions with explicit boundaries. This dimensional transformation allows the system to represent data distributions geometrically rather than statistically, reducing computational complexity while preserving adaptability to different distributions.

Inventive Principle:

Principle #17Another dimension (Dimensionality change)

2Measurement precision

If conventional determination devices use probability density functions to estimate occurrence probabilities, then they can determine categories based on statistical likelihood, but the determination process requires extensive computational resources and time

Engineering Contradiction:

Improvecategory determination accuracyVSAvoiddetermination time

Core Design Contradiction:

Measurement precisionVSLoss of time

Solution Approach 1:

The closed regions are pre-defined in the feature space before actual classification occurs. By establishing the geometric boundaries and node connections in advance, the system eliminates the need for real-time probability density function calculations. The preliminary construction of regions with explicit boundaries enables rapid determination by simply checking which region contains the target data point.

Inventive Principle:

Principle #10Preliminary action

Solution Approach 2:

The patent replaces the statistical-mechanical approach of probability density estimation with a geometric approach. Instead of calculating statistical likelihoods, the system uses spatial relationships and geometric boundaries to determine categories, significantly reducing computational time while maintaining accuracy.

Inventive Principle:

Principle #28Mechanics substitution (Replace mechanical system)

3Device complexity

If conventional determination devices use representative data (prototypes) to represent category distributions, then they can simplify the classification process, but they fail to accurately represent complex data distribution shapes

Engineering Contradiction:

Improveclassification process simplicityVSAvoiddata distribution representation accuracy

Core Design Contradiction:

Device complexityVSMeasurement precision

Solution Approach 1:

Instead of using a single representative prototype, the patent segments each category's data distribution into multiple closed regions defined by connected nodes. This segmentation allows the system to capture complex distribution shapes by creating polyhedral regions that can conform to various geometries, improving representation accuracy while maintaining classification simplicity.

Inventive Principle:

Principle #1Segmentation

Solution Approach 2:

The patent uses closed regions with flexible geometries (not limited to simple spherical or convex shapes) to represent data distributions. By allowing curved and irregular boundaries formed by connecting nodes, the system can accurately represent complex distribution shapes that simple prototypes cannot capture.

Inventive Principle:

Principle #14Spheroidality (Curvature)

Data Source

PatentUS9031882B2Category determination device and method comprising a feature space containing a closed region used to determine the category of a target based on the position of the target within the feature space with reference to the closed region

Publication Date: 2015.05.12 OLYMPUS CORPORATION(JP)

AI summary

A determination device includes a region information recording unit that records therein region information regarding a closed region corresponding to a data distribution shape of a same category within a feature space, the closed region being formed by a plurality of nodes and line segments connecting the plurality of nodes. The determination device also includes a category deciding unit that decides a category of a determination target based on the region information and a position of the determination target within the feature space.