Video Authenticity Detection via Object Focus Variation Analysis
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Surveillance systems face challenges in distinguishing between authentic video captured from a scene and replayed video attacks, where digital signatures may fail to detect videos recorded from displays rather than actual scenes.
Innovation Solution
A method that tracks objects in video frames, measures image quality variations such as focus, pixel contrast, and noise level, and compares these to known variations to determine if the video depicts events from the monitored scene or a display, issuing alerts for potential replayed video attacks.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If digital signatures are applied to captured video, then video transmission authenticity is improved, but detection of replayed video attacks deteriorates
Solution Approach 1:
The patent applies digital signatures to the captured video frames before transmission, embedding authentication data in advance. This preliminary action ensures that the video content itself is verified as originating from the camera sensor, while the present invention adds a separate mechanism to detect replay attacks by analyzing focus characteristics of moving objects in the received video
Solution Approach 2:
The patent introduces focus measure analysis as an intermediary detection mechanism. By measuring the focus characteristics of moving objects in the received video and comparing them against expected focus behavior, the system can detect replay attacks without interfering with the digital signature authentication process. This intermediary analysis layer resolves the contradiction by providing an additional verification dimension
2Measurement precision
If image quality measures are analyzed to detect replay attacks, then detection accuracy is improved, but computational complexity increases
Solution Approach 1:
The patent applies focus measure analysis specifically to regions containing moving objects rather than analyzing the entire video frame. By tracking objects and measuring focus characteristics only in their local regions, the system achieves high detection accuracy while reducing the overall computational burden compared to full-frame analysis
Solution Approach 2:
The patent performs focus measure analysis on a selected subset of video frames containing moving objects rather than analyzing every frame. This partial action approach maintains detection accuracy by focusing computational resources on the most informative frames while reducing overall complexity
Data Source
AI summary
A method for determining authenticity of a video in a surveillance system, whereby a sequence of image frames of a scene is captured, and an object is tracked. A current image quality measure in an image area corresponding to the tracked object is determined in at least a first and second image frame. chosen such that the object has moved at least a predetermined distance between the first and second image frames. A current image quality measure variation for the object is determined, the image quality measure variation describing the image quality measure as a function of position of the object in the image frames. The current image quality measure variation is compared to a known image quality measure variation. In response to the current image quality measure variation deviating from the known pixel density variation by less than a predetermined amount, it is determined that the video is authentic.