Muted Video Call Mouth Obfuscation for NPI Protection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
During video calls or conferences, participants on mute can inadvertently disclose sensitive information through lip movements, which can be discerned by other participants with speechreading capabilities, including uninvited intruders, posing a risk of non-public information leakage.
Innovation Solution
Implementing Artificial Intelligence (AI) with computer vision to monitor mouth movements of muted participants, pausing video capture or transmission, obfuscating the mouth region, or superimposing stationary images to prevent others from viewing speech-related mouth movements, and optionally using Natural Language Processing (NLP) to identify non-public information.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Loss of information
If a call participant places themselves on mute to conduct external conversations, then audio privacy is improved, but visual disclosure of non-public information through lip movements worsens
Solution Approach 1:
The patent extracts the harmful visual information (mouth movements indicating speech) from the video feed by detecting lip movements and selectively blocking or blurring only those regions, while maintaining the rest of the video stream. This allows the system to remove the specific harmful element (visual disclosure of non-public information) without affecting the overall functionality of the video conference.
Solution Approach 2:
The patent introduces an intermediary system consisting of AI-based lip movement detection and NLP analysis that acts as a mediator between the muted participant's external conversations and the other conference participants. This intermediary detects potential non-public information disclosure through visual cues and automatically applies protective measures, enabling the participant to remain muted while preventing information leakage.
2Reliability
If AI-based lip movement monitoring and video blocking are implemented, then non-public information protection is improved, but system complexity worsens
Solution Approach 1:
The patent implements self-service functionality where the system automatically monitors video feeds for lip movements, analyzes speech content using NLP, and applies protective measures without requiring manual intervention from conference participants or administrators. The AI-based system serves itself by detecting potential information disclosure and executing protective actions autonomously, reducing the need for complex manual control mechanisms.
Solution Approach 2:
The patent changes the state of the video feed dynamically based on detected lip movements and speech content. When non-public information is detected, the system modifies video parameters by blocking, blurring, or replacing the affected regions. This parameter-based approach allows the system to adapt to different conversation states and maintain security without requiring fundamental changes to the overall system architecture.
Data Source
AI summary
A system for sensitive data protection in a video call/conference environment. In response to initiating a video call/conference and placing a video call participant on mute, the system uses Artificial Intelligence (AI) specifically computer vision to monitor for mouth movements by the muted video call participant that indicate speech. In response to the monitoring detecting mouth movements that indicate speech, one or more actions are performed that prevent other video call/conference participants from viewing the mouth movement indicating speech by the first call participant. The actions may include stopping/pausing the video feed or capturing of video, obfuscating the region in the video feed that includes the muted video call participant's mouth, or using AI to replace the mouth movements with images of the video call participant's mouth being stationary. The actions may be performed once Non-Public Information (NPI) is identified in the speech.