Video Background Change Characterization for Abandoned Object Detection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing video security systems often fail to reliably distinguish between abandoned and removed objects in scenes, particularly when the background has more texture than the object or when contours do not accurately match, leading to incorrect conclusions.
Innovation Solution
The system performs an image subtraction operation between current frames and a background model to identify changed pixel patches, analyzes edge strengths in both models, and characterizes objects as abandoned or removed based on the comparison of edge strengths, without requiring accurate segmentation or assuming relative edge densities.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If the analysis system compares contours of changed pixel patches to edges detected in masked areas, then object classification can be performed, but the system makes mistakes when the background has similar features as foreground objects
Solution Approach 1:
The patent divides the image into changed pixel patches and applies masking to isolate specific regions. By segmenting the image and focusing analysis on changed areas rather than the entire scene, the system reduces interference from background features that resemble foreground objects, thereby improving classification reliability
Solution Approach 2:
The patent applies different analysis methods to different regions: changed pixel patches are analyzed separately from the rest of the image using masking. This local quality approach allows the system to focus computational resources on areas with actual changes while ignoring distracting background features, improving both accuracy and reliability
2Measurement precision
If the analysis system measures similarity of content inside and outside changed pixel patches, then object presence can be detected, but the system makes mistakes when contours do not accurately match contours of the object
Solution Approach 1:
Instead of requiring complete contour matching, the patent measures similarity of content inside and outside the changed pixel patches. This partial action approach detects objects based on content similarity in the changed region alone, eliminating the need for precise contour alignment and reducing errors from segmentation inaccuracies
3Measurement precision
If the analysis system detects edges in changed areas and compares edge counts, then abandoned vs removed objects can be distinguished, but the system makes mistakes when the background has more texture than the object
Solution Approach 1:
The patent extracts and isolates changed pixel patches from the rest of the image using masking techniques. By taking out only the changed regions for analysis, the system eliminates background texture from the comparison process, allowing accurate distinction between abandoned and removed objects regardless of background complexity
Data Source
AI summary
A method and system for analyzing video data in a security system. An analysis compares a current frame to a background model. The analysis system compares the background model to the current frame to identify changed pixel patches. The analysis system uses morphological image processing to generate masks based on the changed pixel patches. Next, the analysis system applies the masks to the background model and the current frames to determine whether the changed pixel patches are characteristic of abandoned or removed objects within the video data.