Foreground Detection Accuracy in Video Surveillance Systems
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing foreground detection methods in image processing suffer from false detection due to background movements and low accuracy in graph segmentation algorithms, leading to incorrect classification of visual elements as foreground or background.
Innovation Solution
An image processing apparatus that acquires a current image and a background model, classifies visual elements, determines similarity measures between groups in the current and background images, and identifies falsely classified elements by referencing previous foreground classifications, thereby improving detection accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If background subtraction techniques are used to classify visual elements as foreground or background, then foreground detection can be performed, but false foreground detection occurs when background elements (water ripples, leaves) change visually
Solution Approach 1:
The patent applies preliminary action by pre-classifying visual elements into foreground and background categories before actual detection. A background model is constructed in advance containing classified visual elements, which serves as a reference for subsequent detection. This pre-classification allows the system to distinguish between actual foreground objects and background elements that may undergo visual changes, thereby reducing false detections while maintaining detection accuracy.
2Productivity
If graph segmentation algorithm with low accuracy is used to obtain super-pixels, then processing speed is maintained, but visual elements cannot keep constant with previous images leading to false detection
Solution Approach 1:
The patent applies beforehand cushioning by creating a background model that stores classified visual elements from previous images. This background model acts as a cushion or buffer that compensates for the low accuracy of graph segmentation algorithms. When detecting foreground elements, the system references this pre-established background model, which cushions against the inconsistencies and errors introduced by low-accuracy segmentation, thereby maintaining visual element consistency without sacrificing processing speed.
3Ease of operation
If background confidence is calculated based on neighboring visual elements, then classification can be performed, but when neighboring elements are falsely detected as foreground, the background confidence becomes smaller and less than threshold
Solution Approach 1:
The patent applies feedback by using the classified visual elements from the background model to verify and adjust the classification of current visual elements. The system continuously compares current frame classifications against the established background model, and this feedback loop allows the system to correct false detections. When neighboring elements are incorrectly classified, the feedback mechanism references the background model to re-evaluate and correct the classification, maintaining accurate background confidence values.
Data Source
AI summary
Acquiring a current image from an inputted video and a background model which comprises a background image and foreground/background classification information of visual elements; classifying the visual elements in the current image as foreground or background; determining similarity measures between the current image and groups in the background model, wherein visual elements in the current image are the visual elements in the current image which are classified as the foreground, wherein visual elements in the groups in the background model are the visual elements whose classification information is the foreground, and wherein the visual elements in the groups in the background model are the visual elements which neighbour to corresponding portions of the visual elements in the groups in the current image; and identifying whether the visual elements in the current image which are classified as the foreground are falsely classified or not according to the determined similarity measures.