Dynamic Mask Template Matching for Deformed Target Detection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing image detection methods using template matching struggle with deformation of surveillance targets, leading to difficulties in tracking and potential intrusion of background into the template, which affects detection accuracy.
Innovation Solution
An image detection device and method that generates a template and a dynamic mask based on temporal variations of feature points, optimizing the template shape to improve deformation handling and reduce background interference during target detection.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If template matching is used for target detection, then high matching performance and ability to track target again after switching or leaving surveillance area are achieved, but the method becomes vulnerable to deformation of surveillance target
Solution Approach 1:
The patent applies dynamics by making the template shape adaptive and changeable over time. The template generation unit creates an initial template, and the shape optimization unit continuously refines the template shape based on feature point distribution from tracking results. This dynamic adaptation allows the template to maintain high matching performance while accommodating target deformation, resolving the contradiction between tracking reliability and deformation adaptability.
Solution Approach 2:
The patent changes the shape parameters of the template based on feature point distribution. The shape optimization unit adjusts template boundaries and geometry according to the spatial distribution of tracked feature points, allowing the template to morphologically adapt to target deformation while maintaining reliable tracking through iterative parameter optimization.
2Measurement precision
If template size is optimized to exclude background, then background intrusion is reduced, but the template may not accommodate deformed target shapes
Solution Approach 1:
The template shape is made dynamic through continuous optimization based on feature point distribution. Rather than using a fixed optimized size, the template adapts its geometry iteratively to match the deformed target shape while maintaining precise boundaries that exclude background, resolving the contradiction between detection precision and shape adaptability.
Solution Approach 2:
The patent applies local quality by optimizing different regions of the template based on local feature point distribution. The shape optimization unit adjusts specific portions of the template boundary according to where feature points are concentrated, allowing precise fitting to the target shape while maintaining exclusion of background regions, thus achieving both detection precision and shape adaptability.
3Adaptability or versatility
If feature points are used for tracking, then robustness against target deformation is achieved, but the right target cannot be tracked again after switching or leaving surveillance area
Solution Approach 1:
The patent segments the tracking approach into two components: feature point-based tracking for robust deformation handling, and template matching for reliable target identification. The system uses both methods in combination, where feature points provide deformation robustness and template matching ensures correct target reidentification, resolving the contradiction between deformation robustness and tracking reliability.
Solution Approach 2:
The patent merges feature point tracking with template matching into a unified detection system. The feature point distribution informs template shape optimization, while the template provides a reference for verifying target identity. This combination achieves both deformation robustness from feature points and reliable retracking from template matching.
Data Source
AI summary
Provided are an image detection device, an image detection method and a program, which are capable of improving correspondence to a target deformation by optimizing a template shape, when performing target detection using template matching. An image detection device 100 for detecting a target from an input image comprises: a template generation unit 10 that generates a template for detecting a target; a mask generation unit 20 that generates a mask which shields a portion of the template, on the basis of temporal variations of a feature point extracted from an area including the image target; and a detection unit 30 that detects the target from the image using the template a portion of which is shielded by the mask.