Surveillance Video Privacy Control via Dynamic ROI Masking
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing video surveillance systems face limitations in providing privacy control, including reduced compression ratio, increased bandwidth requirements, poor video quality, and limited dynamic privacy levels, due to the camera identifying and encrypting regions of interest, which leads to high complexity and resource consumption.
Innovation Solution
A computing device identifies and masks regions of interest in surveillance video based on user-provided authentication keys, allowing for dynamic and adaptive privacy control without increasing bandwidth or resource demands on the camera, and maintaining video quality by processing on the computing device rather than the camera.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If ROIs are encrypted at the camera end, then privacy control is provided, but compression ratio is reduced and bandwidth requirement increases
Solution Approach 1:
The patent extracts the ROI identification and encryption functions from the camera end to a separate processing system. The camera only captures and transmits the full video stream, while ROI identification, authentication level determination, and selective encryption are performed by an external processing system, thereby avoiding the bandwidth overhead of transmitting multiple encrypted streams from the camera.
Solution Approach 2:
The system performs preliminary authentication and determines the user's authentication level before providing video access. Based on this pre-determined authentication level, the system dynamically identifies appropriate ROIs and applies encryption only to those regions, rather than encrypting all regions or requiring multiple pre-encoded streams from the camera.
2Adaptability or versatility
If multiple encoded streams are generated for different ROIs, then different privacy levels are supported, but complexity and power consumption at the camera end increases exponentially
Solution Approach 1:
The patent extracts the complex ROI identification and multi-stream encoding functions from the camera to an external processing system. The camera's role is simplified to only capturing and transmitting the full video stream, while all complexity-related operations (authentication, dynamic ROI identification, selective encryption) are performed externally, reducing camera complexity to minimal levels.
Solution Approach 2:
The system implements dynamic ROI identification based on real-time authentication levels. Instead of the camera generating fixed multiple streams, the external processing system dynamically determines which regions to encrypt based on the user's authentication level, allowing flexible adaptability without increasing camera complexity.
3Reliability
If ROIs are encrypted during encoding, then privacy control is provided, but decoded video quality is reduced
Solution Approach 1:
The patent introduces an intermediary processing system that receives the full decoded video stream, identifies ROIs, and applies encryption selectively to those regions after decoding. This intermediary approach allows the video to be fully decoded first (maintaining quality), then selectively encrypted for privacy control, rather than encrypting during the encoding process which degrades quality.
Data Source
AI summary
Method and a computing device for providing privacy control in a surveillance video is disclosed. The computing device receives a surveillance video from a video capturing. The computing device also receives an authentication input key from a user of the computing device. Based on the authentication input key, the computing device identifies one or more region of interests (ROIs) in the video surveillance. The identified ROIs are masked by the computing device for providing the privacy control. In the present disclosure, the number of privacy levels achievable are unlimited and not fixed by ROIs at the video capturing device end. The increase in number of privacy levels do not require any changes or complexity/power increase in the video capturing device end and doesn't demand for increase in bandwidth from the video capturing device. Further, it doesn't deteriorate the video quality.