AR Occlusion for Privacy Backgrounds in Live Video
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing video recording methods are unreliable in effectively occluding privacy background images during live broadcasts, especially when the camera moves or the background is dynamic, leading to incomplete or ineffective concealment of unwanted objects.
Innovation Solution
A video recording method that acquires sensitive video frames, identifies privacy background images, and superimposes augmented reality (AR) occlusion images to occlude these images, using dynamic or static background type analysis to determine the appropriate occlusion image, ensuring real-time and reliable hiding of sensitive content.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If existing manual occlusion methods are used during live broadcast, then the anchor can control the occlusion timing, but the occlusion is unreliable when the camera moves or background is dynamic
Solution Approach 1:
The system performs automatic occlusion processing without requiring manual intervention. The processor automatically identifies privacy background images in video frames and applies occlusion images to conceal them, enabling the system to serve itself in detecting and occluding sensitive content throughout the live broadcast.
Solution Approach 2:
The system continuously monitors video frames during live broadcast, detects privacy background images in real-time, and dynamically adjusts occlusion application based on the detected content. This closed-loop feedback mechanism ensures reliable occlusion even when camera movement or dynamic backgrounds occur.
2Adaptability or versatility
If the camera moves during live broadcast, then the anchor can capture different backgrounds, but the existing occlusion method cannot track and occlude moving privacy objects
Solution Approach 1:
The system pre-processes video frames to identify privacy background images before final occlusion application. By detecting and marking sensitive regions in advance within each frame, the system prepares occlusion data that can be reliably applied even when the camera moves or backgrounds change dynamically during broadcast.
3Productivity
If artificial occlusion is performed timely during live broadcast, then the anchor can attempt to conceal sensitive content, but the occlusion cannot be performed effectively when the lens moves with the anchor
Solution Approach 1:
The system automatically performs occlusion processing without requiring manual timing or intervention. The processor continuously analyzes video frames, identifies privacy background images, and applies occlusion images automatically, eliminating the need for manual timing while maintaining both speed and accuracy of occlusion.
Data Source
AI summary
A video recording method and apparatus, an electronic device and a readable storage medium, which realize positioning of a privacy background by acquiring a sensitive video frame containing a privacy background image from the recorded video information; acquire an image region of the privacy background image in the sensitive video frame; superimpose an augmented reality AR occlusion image on the image region of the privacy background image to obtain an occlusion frame in which the privacy background image is occluded, so as to realize the hiding of the privacy background image by occluding part of the background; obtain live video information according to the recorded video information and the occlusion frame, thereby improving the reliability of real time occlusion of objects that need to be hidden, during video recording.