Object Recognition Apparatus Segmentation for Accuracy
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing object recognition methods integrate multiple information properties to calculate a feature amount that averages both advantageous and disadvantageous information, leading to a ceiling in recognition accuracy.
Innovation Solution
An object recognition apparatus that generates property data highlighting specific properties, extracts discrimination-use and reliability feature amounts, calculates discrimination and reliability information, and synthesizes this information for accurate object recognition.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If multiple information properties are integrated to calculate a feature amount, then comprehensive object representation is achieved, but recognition accuracy reaches a ceiling due to averaging advantageous and disadvantageous information
Solution Approach 1:
The patent segments the integrated information into multiple pieces of information, each associated with different properties. Instead of calculating a single averaged feature amount from integrated information, the system extracts multiple feature amounts separately from each piece of information, preserving the distinctive characteristics of each property without mutual contamination or averaging.
2Adaptability or versatility
If multiple information properties are integrated into one piece of information, then comprehensive representation is achieved, but discrimination capability is reduced due to information averaging
Solution Approach 1:
The patent maintains segmentation of information by property type throughout the processing pipeline. Each piece of information with different properties is kept separate during feature extraction, and the final synthesized information combines multiple distinct feature amounts rather than a single averaged feature, preserving discrimination capability while achieving comprehensive representation.
Solution Approach 2:
The patent creates a composite information structure where multiple feature amounts (each derived from different property types) are synthesized together. This composite approach allows the system to leverage the strengths of each property type while maintaining the discriminatory power of individual features, avoiding the degradation that occurs with simple averaging.
Data Source
AI summary
Provided is technique of object recognition that can accurately recognize an object. An object recognition apparatus (i) generates property data that highlights a specific property based on target data, (ii) extracts a discrimination-use feature amount used for discrimination of each piece of the property data, (iii) calculates discrimination information used for discrimination of the property data, (iv) extracts a reliability feature amount used for estimation of reliability of the discrimination information calculated for each piece of the property data, (v) estimates the reliability of the discrimination information, (vi) generates synthesized information acquired by synthesizing the discrimination information calculated for each piece of the property data and the reliability calculated for each piece of the property data, and (vii) performs processing related to the object recognition.