Target Extraction Background Region Adjustment
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing target extraction methods fail to reliably separate targets from backgrounds due to the lack of criteria for setting the background region, leading to deteriorated tracking performance and noise sensitivity.
Innovation Solution
The method adjusts the size and distance of the background region based on the target region, using a weight value function to increase the number of pixels and generate histograms, thereby effectively separating the target from the background and improving tracking reliability.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If the background region is set without specific criteria, then the target extraction process is simple, but the tracking reliability deteriorates due to inability to obtain reliable statistical information
Solution Approach 1:
The patent applies parameter changes by establishing specific criteria for background region setting, including distance from target center and size proportionality. The background region is defined as an area where distance to target center is greater than a first threshold and area ratio exceeds a second threshold, transforming the vague concept of background region into precisely controlled parameters that ensure reliable statistical information for tracking.
2Reliability
If the background region size is too small, then the processing speed is fast, but the statistical characteristics become unreliable leading to poor target extraction
Solution Approach 1:
The patent implements dynamics by making the background region size adaptive rather than fixed. The background region dynamically adjusts its size based on the target region size, maintaining a proportional relationship where the background region area is 0.5 to 2 times the target region area. This dynamic adjustment ensures sufficient statistical samples while optimizing processing efficiency for different target scales.
3Measurement precision
If the background region includes target pixels, then the coverage is complete, but the target and background separation becomes inaccurate
Solution Approach 1:
The patent applies local quality by creating spatially differentiated regions with distinct characteristics. The target region is defined as an circular area centered on the detected target, while the background region is specifically positioned outside this target region at a distance greater than a first threshold. This spatial differentiation ensures that background pixels are truly representative of the background and do not contaminate target statistics, achieving accurate separation while maintaining sufficient background coverage.
Data Source
AI summary
Provided are an apparatus and a method for extracting a target related to an algorithm separating the target and a background by using statistical characteristics of the target and the background among target extracting methods required in a weight value center tracking method, and recording media storing a program performing the method. According to the present invention, it is possible to effectively separate the target region and the background region from each other and it is possible to improve reliability in target extracting performance.