Collaborative Session Recording via Image Segmentation and Static Extraction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Recording collaborative audio-visual sessions is challenging due to high memory and computing resource requirements, especially when compressing real-time video, which often results in low frame rates and inefficient data storage.
Innovation Solution
A two-phase system and method for recording collaborative sessions, where the first phase occurs in real-time by identifying and removing constant or preprocessed components from the composite image, allowing for efficient compression, and the second phase reconstructs the image offline using element streams and audio, optimizing CPU resources and achieving high-quality data compression.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If standard encoding techniques (e.g., H.264) are used to compress the entire composite image in real-time, then video quality can be maintained, but CPU resources are excessively consumed and frame rate drops to only 2 fps
Solution Approach 1:
The patent segments the composite image into multiple layers including background layer, foreground video layers, and annotation layers. Each layer is processed independently with appropriate compression techniques, avoiding the need to compress the entire high-resolution composite image as a single unit, thereby reducing CPU load while maintaining video quality
Solution Approach 2:
The patent extracts and removes constant or static portions of the composite image (such as static backgrounds or unchanged regions) before compression. By taking out these redundant elements, the amount of data requiring real-time compression is significantly reduced, enabling higher frame rates without sacrificing quality in the dynamic regions
2Loss of information
If the entire composite image is compressed in real-time, then complete recording is achieved, but memory requirements and computing resources become unmanageably large
Solution Approach 1:
The recording system segments the composite image into multiple layers (background, foreground videos, annotations) and processes each layer separately. This segmentation allows selective compression and storage of only the essential dynamic content, significantly reducing memory and storage requirements while preserving all important collaborative information
Solution Approach 2:
The system extracts and removes constant or static portions of the composite image that do not change during the collaborative session. By eliminating these redundant static elements from real-time compression and storage, the system achieves complete recording of dynamic content with minimal memory and storage overhead
3Productivity
If high frame rate recording (15-30 fps) is attempted with standard compression, then smooth playback is achieved, but CPU resources are excessively consumed
Solution Approach 1:
The patent divides the composite image into multiple processable layers and independently compresses only the dynamic foreground elements at high frame rates, while static backgrounds are handled separately or removed. This segmentation enables smooth high-frame-rate playback of collaborative content without the CPU overhead of compressing the entire high-resolution composite image
Solution Approach 2:
The system extracts and removes constant or static portions of the composite image before real-time compression. By eliminating these redundant static elements, the amount of data requiring high-speed compression is dramatically reduced, enabling 15-30 fps frame rates with minimal CPU resource consumption
Data Source
AI summary
A system and method for recording a collaborative session includes two phases. One is performed in real-time and includes determining portions of a composite image of collaborative session content that are constant over time or preprocessed. The portions are removed from the composite image of the collaborative session content. Remaining content of the composite image and any of the portions not already compressed is compressed and stored. A second phase is performed when off-line and includes decoding the remaining content of the composite image and the portions to reconstruct the composite image of the collaborative session content and formatting the composite image.