Feature Point Selection for Accurate Target Detection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing image processing techniques fail to accurately detect the position and posture of targets in images due to the lack of consideration for feature points, leading to erroneous detection, especially in complex shapes and when distinguishing between similar targets.
Innovation Solution
An image processing device that uses a frequency calculation unit to determine notable feature points based on the comparison of standard shape information defined by multiple feature points, allowing for accurate detection by distinguishing between correct and incorrect detection results.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If matching is performed using all feature points in the model pattern, then the detection process can be completed, but detection accuracy deteriorates due to inclusion of non-notable feature points that do not contribute to accurate target identification
Solution Approach 1:
The patent extracts only the notable feature points from the model pattern that contribute to accurate target detection. The notable feature point extraction unit identifies and extracts specific feature points based on frequency of appearance and contribution to detection accuracy, excluding non-notable feature points that do not help in accurate identification. This extraction principle resolves the contradiction by selecting only the essential subset of feature points.
Solution Approach 2:
The patent applies local quality by treating different feature points differently based on their importance. Notable feature points are identified and given special attention through frequency analysis and contribution evaluation, while non-notable feature points are excluded. This differential treatment based on local importance resolves the contradiction between using all feature points and achieving accurate detection.
2Measurement precision
If manual selection of notable regions is performed, then detection accuracy can be improved, but human error increases leading to incorrect region selection
Solution Approach 1:
The patent implements self-service by enabling the system to automatically extract notable feature points without human intervention. The notable feature point extraction unit performs frequency analysis and contribution evaluation autonomously to identify important feature points. This automation eliminates human error in region selection while maintaining high detection accuracy through objective computational criteria.
Solution Approach 2:
The patent changes the parameter of feature point selection from manual visual inspection to automated frequency-based analysis. By transforming the selection criterion into a quantifiable parameter (frequency of appearance and contribution to detection), the system achieves both high accuracy and reliability without human error. This parameter transformation resolves the contradiction between manual selection accuracy and reliability.
3Measurement precision
If feature points are selected based on frequency analysis, then notable feature points can be identified objectively, but the processing time increases
Solution Approach 1:
The patent applies preliminary action by pre-calculating and storing the frequency of appearance and contribution values for each feature point during the model pattern creation phase. These pre-computed parameters are then reused during actual detection without recalculation, significantly reducing processing time while maintaining accurate feature point selection. This preliminary computation resolves the contradiction between analysis accuracy and processing time.
Solution Approach 2:
The patent performs partial action by selecting only the top notable feature points based on frequency and contribution thresholds, rather than analyzing all possible feature points equally. This selective approach reduces the number of feature points requiring detailed analysis during detection, thereby reducing processing time while maintaining selection accuracy through focused evaluation of the most important features.
Data Source
AI summary
An image processing device comprises: a result acquisition unit that acquires one or more of the input images including a target, and acquires a detection result obtained by comparing feature points of standard shape information with input-side feature points extracted from the input image; a frequency calculation unit that acquires multiple detection results in which the standard shape information and the target are placed in different positions and different postures, and calculates frequencies of detection of the input-side feature points in the input image for corresponding ones of the feature points of the standard shape information; and a feature point selection unit that selects a notable feature point from the feature points of the standard shape information on the basis of the frequencies calculated by the frequency calculation unit.