Image Detection Using Characteristic Point ROIs
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional image detection techniques require a full-field scan, leading to inefficiency due to a large number of scan windows and detection of non-target images, despite advancements like the Speeded Up Robust Features (SURF) algorithm, which still suffers from inefficiency due to block scans.
Innovation Solution
The method employs an importance sampling algorithm based on image characteristic points to reduce the number of sampling regions by creating regions of interest (ROIs) and using SVM for target hit score calculations, along with a centroid shift weight equation to narrow the search scope, thereby enhancing detection efficiency without compromising accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If a full-field scan is performed on the entire picture to detect target images, then detection accuracy is maintained, but detection efficiency deteriorates due to the large number of scan windows required
Solution Approach 1:
The image is segmented into multiple regions based on characteristic points extracted by SURF algorithm. Instead of scanning the entire image uniformly, the method divides the image into relevant regions (ROIs) centered at characteristic points, and further segments these ROIs into sub-regions for targeted detection. This segmentation reduces the number of scan windows from covering the full image to only covering important regions, thereby improving detection efficiency while maintaining accuracy.
Solution Approach 2:
The method applies different detection strategies to different regions of the image. High-priority regions (those containing characteristic points) receive focused detection with multiple scan windows, while low-priority regions are either skipped or given minimal attention. This local quality approach ensures that detection resources are concentrated where they are most needed, improving overall efficiency without sacrificing detection accuracy in critical areas.
2Speed
If the SURF algorithm is used to retrieve characteristic points and perform block scan, then detection speed is improved, but detection efficiency deteriorates due to the presence of non-target image blocks in the scan
Solution Approach 1:
The method extracts and isolates regions containing characteristic points from the rest of the image. By creating ROIs centered at characteristic points and then further segmenting these into sub-regions, the method takes out only the relevant portions of the image that need detailed scanning. Non-target blocks are excluded from intensive scanning, thereby improving detection efficiency while maintaining the speed advantage of SURF-based characteristic point retrieval.
Solution Approach 2:
Instead of performing uniform block scans across the entire image or even across all characteristic point regions, the method applies partial scanning only to specific sub-regions within ROIs. The scan windows are strategically placed only where targets are most likely to be found, based on characteristic point locations. This partial action approach eliminates wasted scanning of non-target areas while maintaining detection speed.
3Productivity
If the number of scan windows is reduced by using characteristic points, then detection efficiency is improved, but detection accuracy may deteriorate due to fewer sampling regions
Solution Approach 1:
The method adds a hierarchical dimension to the detection process. Instead of a single-level scan window arrangement, it creates a two-level structure: first level defines ROIs around characteristic points, and the second level subdivides these ROIs into multiple sub-regions with scan windows. This dimensional change allows the system to concentrate more scan windows in critical areas (increasing local sampling density) while reducing the overall number of scan windows across the entire image, thereby maintaining accuracy while improving efficiency.
Data Source
AI summary
A method of speeding up image detection, adapted to increase a speed of detecting a target image and enhance efficiency of image detection, comprises the steps of capturing an image; retrieving a plurality of characteristic points of the image; creating a region of interest (ROI) centered at the characteristic points each; creating a plurality of search point scan windows corresponding to the ROIs, respectively; calculating target hit scores of the characteristic points and the search point scan windows; comparing the target hit scores of the characteristic points and the search point scan windows to obtain an ROI most likely to have a target image; calculating centroid coordinates of the ROI by a centroid shift weight equation; and narrowing a scope of ROI search according to a location of the centroid coordinates and reducing a displacement between the search points.