Virtual Object Video Call Interface for Privacy Protection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
In existing video call session technologies, users may not want to participate in video calls using their real images.
Innovation Solution
A method for video call sessions that involves displaying a virtual object corresponding to the user in the video call interface, recognizing the user's movements from a collected image, and controlling the virtual object to perform these movements, thereby allowing users to interact without showing their real images.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If users participate in video call sessions with real images, then communication authenticity is improved, but user privacy and comfort deteriorate
Solution Approach 1:
The patent creates virtual objects that copy and simulate the user's real-world movements, expressions, and gestures. The virtual object serves as a digital representation that replicates the user's actions in real-time, allowing authentic interaction without exposing the user's actual image. This copying mechanism resolves the contradiction by maintaining communication authenticity through accurate movement replication while protecting privacy by never displaying the real user image.
Solution Approach 2:
The virtual object acts as an intermediary between the user and the video call interface. Instead of directly displaying the user's real image, the system uses the virtual object as a mediator that represents the user's presence and actions. This intermediary layer preserves the user's privacy while maintaining the authenticity of communication by accurately reflecting the user's movements and expressions through the virtual representation.
2Ease of operation
If virtual objects are used to protect user privacy, then user comfort is improved, but system complexity increases
Solution Approach 1:
The patent replaces traditional mechanical video capture and display systems with an AI-driven virtual object generation system. Instead of using cameras and direct video transmission, the system uses image recognition technology to capture user movements and AI algorithms to generate corresponding virtual object animations. This substitution reduces the complexity of hardware requirements while improving user comfort through privacy protection.
Solution Approach 2:
The system automatically performs image recognition, movement analysis, and virtual object generation without requiring manual user intervention. The AI algorithms self-service by automatically tracking user movements, interpreting gestures, and animating the virtual object in real-time. This automation reduces the operational burden on users while managing system complexity through integrated AI processing.
3Speed
If real-time image recognition is implemented, then virtual object responsiveness is improved, but processing time increases
Solution Approach 1:
The system performs preliminary actions by pre-processing and pre-training AI models for efficient real-time inference. Image recognition algorithms are pre-trained on extensive datasets to enable rapid movement detection during actual video calls. Key movement patterns and expressions are pre-analyzed and categorized, allowing the system to quickly match recognized movements with corresponding virtual object animations without extensive real-time processing.
Solution Approach 2:
The system dynamically adjusts processing intensity based on the complexity of detected movements. For simple gestures, the system uses lightweight recognition models that process quickly, while more complex expressions trigger more detailed analysis. The virtual object animation complexity also dynamically adapts to match the detected movement intensity, optimizing the balance between responsiveness and processing time in real-time.
Data Source
AI summary
Embodiments of this application disclose a method for video call session and apparatus, an electronic device, a storage medium, and a program product. The method includes displaying a first video call session interface on a first client participating in a video call session, a first virtual object corresponding to the first client in the first video call session interface; recognizing a first movement of a first user in a first image, the first image being an image collected by the first terminal when the first terminal faces the first user; and controlling the first virtual object in the first video call session interface to perform the first movement.