Object Recognition via Viewing Direction Weighted Feature Matching
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current monitoring systems with multiple cameras face challenges in efficiently acquiring and matching images of the same object across different cameras, often requiring manual effort and being time-consuming, especially when objects are viewed from varying angles.
Innovation Solution
An object recognizing apparatus and method that estimates the viewing direction of objects captured by different cameras, extracts features, and allocates weights based on these directions to calculate similarity, determining if objects are the same, thereby improving object recognition accuracy and efficiency.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If manual matching of images from multiple cameras is performed, then object recognition can be achieved, but time consumption and human effort increase significantly
Solution Approach 1:
The patent replaces the manual mechanical matching process with an automated computer-based system that uses viewing direction estimation and feature weighting algorithms to automatically match objects across multiple camera images, eliminating human labor and significantly reducing time consumption while maintaining recognition accuracy
2Productivity
If simple feature matching based on color and texture is used, then processing speed is fast, but recognition accuracy deteriorates when objects appear from different viewing directions
Solution Approach 1:
The patent applies local quality by assigning different weights to different features based on their reliability under specific viewing conditions. The system estimates viewing directions and adjusts feature weights locally according to each feature's suitability for that particular viewing angle, thereby maintaining high recognition accuracy across diverse perspectives without sacrificing processing speed
Solution Approach 2:
The patent dynamically changes the weighting parameters of features based on estimated viewing directions. By adjusting the importance of different features (color, texture, shape) according to the camera's viewing angle relative to the object, the system adapts to different viewing conditions and maintains high recognition accuracy while preserving fast processing speeds
3Device complexity
If uniform weighting is applied to all features, then calculation is simple, but recognition accuracy decreases when background differences and varying viewing angles are present
Solution Approach 1:
The patent introduces dynamics by making feature weights adaptive rather than static. The system continuously estimates viewing directions and adjusts feature weights dynamically based on the current viewing conditions, allowing the matching process to adapt to varying backgrounds and angles without requiring overly complex fixed-weight schemes
Data Source
AI summary
An object recognizing apparatus and method are provided. The apparatus may include: a viewing direction estimating device configured for respectively estimating a first viewing direction of a first object captured by a first camera and a second viewing direction of a second object captured by a second camera; a feature extracting device configured for extracting one or more features respectively from an image containing the first object captured by the first camera and an image containing the second object captured by the second camera; and an object matching device configured for allocating a weight for each of the one or more features according to the first viewing direction and the second viewing direction, and calculating a similarity between the first object and the second object based on the one or more weighted features, to determine whether the first object and the second object are the same object.