3D Model Search Accuracy via Intensity-Corrected Feature Vectors
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current three-dimensional model search methods, such as DSR 472, do not achieve the highest search accuracy due to limitations in feature vector computation, particularly in emphasizing pixel gray scales in two-dimensional projected images.
Innovation Solution
A method that corrects two-dimensional projected images by emphasizing pixel gray scales as the distance from the center increases, performs a Fourier transform, and computes feature vectors from the intensity-modified images, improving search accuracy by using low-frequency components for similarity comparison.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If conventional feature vector computation methods are used without intensity correction, then the computation process is simple, but the search accuracy is insufficient
Solution Approach 1:
The patent applies preliminary action by performing intensity correction on the projected image before computing the feature vector. The correction emphasizes pixel intensities based on their distance from the image center, creating a pre-processed image that enhances contour features. This preliminary processing step improves search accuracy without significantly increasing overall system complexity.
Solution Approach 2:
The patent changes the intensity parameter of pixels in the projected image based on their position. Specifically, it applies a correction function that modifies pixel intensities according to their distance from the center, transforming the original image into an intensity-corrected version that emphasizes edge and contour information, thereby improving feature extraction quality.
2Measurement precision
If pixel gray scales are emphasized as distance from center increases, then contour features are enhanced and search accuracy improves, but the image processing complexity increases
Solution Approach 1:
The patent applies local quality by differentiating the treatment of different regions in the image. Pixels closer to the center maintain their original intensity, while pixels farther from the center have their intensities emphasized. This localized processing enhances contour features at the periphery without unnecessarily processing the central region, optimizing the balance between feature enhancement and processing complexity.
Applied Scientific Principles
This section explains which scientific principles are used to turn an abstract innovation direction into a practical engineering solution.
Function Achieved in This Case
This approach enhances search accuracy by emphasizing contour features and improving similarity search performance, outperforming conventional methods in terms of precision and recall.
Implementation Method 1
performing a Fourier transform of the intensity-modified two-dimensional image; computing a feature value from the result of the Fourier transform
Data Source
AI summary
A novel search method having an advantage over the DSR 472 method the search accuracy of which is currently said to be the highest level is proposed. At the feature value computation stage, the two-dimensional image used in computation of the three-dimensional model feature value (excluding a contour feature value) is corrected so that the gray scales of the pixels are emphasized more as the distance from the center of the two-dimensional image increases.