Generative AI Text Generation From Changed Screen Frames
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing video conference systems struggle to effectively capture and summarize text-based content shared during meetings, making it difficult to identify or search for specific meeting content later, especially when security is a concern.
Innovation Solution
Utilizing generative AI to automatically generate meeting minutes, action items, and questions/answers by extracting and combining text from shared screens, recognizing changes in the screen content, and integrating speech data to improve accuracy and security monitoring.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If screen sharing is performed by capturing and compressing video using a video codec, then various types of content can be shared freely, but it becomes difficult to check or search for specific text-based content later
Solution Approach 1:
The patent extracts text from shared screen content using OCR technology, separating the text information from the video stream. This extracted text is then stored independently, allowing users to search and review specific text content without needing to watch the entire video recording, thus resolving the contradiction between versatile content sharing and text accessibility.
2Reliability
If video recording is used to check shared content later, then content can be reviewed, but it requires playing back the entire video which is time-consuming and inefficient
Solution Approach 1:
The patent extracts and stores text content separately from the video recording. This allows users to directly search and review extracted text instead of playing back entire video recordings, significantly reducing the time required to verify shared content while maintaining reliability through accurate OCR extraction.
Solution Approach 2:
The patent creates a text copy of the shared screen content through OCR extraction. This text copy serves as an independent reference that can be searched and reviewed without accessing the original video, enabling efficient content verification without time-consuming video playback.
3Productivity
If only video compression is used for screen sharing, then transmission is efficient, but security monitoring of shared content becomes difficult
Solution Approach 1:
The patent extracts text from shared screen content and stores it separately from the video stream. This extracted text enables security officers to efficiently monitor and search for sensitive information without needing to review entire video recordings, thus improving security monitoring capability while maintaining transmission efficiency through continued use of video compression.
Data Source
AI summary
The disclosure relates to a method, an apparatus, and a computer program for automatically generating text using generative AI. A method for automatically generating text using generative AI of the disclosure may include determining whether or not there is a change between frames constituting images; extracting text from a changed frame; and generating first text by combining the extracted text with previously extracted text.