Triggered Image Description Text Distribution for Surveillance Events
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing network cameras lack the ability to convert image content into text descriptions and push important surveillance information to user terminals in real-time, leading to inefficiencies in information retrieval due to bandwidth limitations and excessive data processing.
Innovation Solution
A method and system for generating textual descriptions of images or videos based on predefined keywords and important information, using image recognition and keyword generation models to distribute relevant text to user terminals only when specific triggering conditions are met.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Loss of information
If every image content is converted into text descriptions and pushed to user terminals, then users can obtain complete surveillance information, but it results in very large amount of image processing computation and users receiving unwanted information
Solution Approach 1:
The patent extracts only the important information from images by performing image recognition to identify objects and events that satisfy predetermined triggering conditions. Instead of converting all image content to text, the system selectively extracts and processes only relevant portions, thereby reducing computation while maintaining information quality.
Solution Approach 2:
The patent applies different processing quality levels to different parts of the image stream. Important images containing objects or events of interest receive full text description processing, while other images are skipped or processed minimally. This local differentiation optimizes resource allocation based on actual information value.
2Loss of information
If every image content is converted into text descriptions and pushed to user terminals, then users can obtain complete surveillance information, but it leads to users receiving unwanted information
Solution Approach 1:
The system extracts only important information by identifying objects and events that meet predetermined triggering conditions. This selective extraction filters out unwanted information while preserving essential surveillance data, solving the contradiction between completeness and information quality.
Solution Approach 2:
The system uses image recognition results as feedback to determine whether an image contains important information. This feedback mechanism enables intelligent filtering, where only images with recognized objects or events satisfying triggering conditions are converted to text and pushed to users, eliminating unwanted information推送.
3Loss of time
If image content is converted into text descriptions in real-time, then users can quickly obtain important surveillance information, but it requires very large amount of image processing computation
Solution Approach 1:
The patent extracts only important images for real-time text conversion by applying image recognition and triggering conditions. This selective approach reduces the volume of images requiring computation while maintaining real-time delivery of critical information, balancing speed and computational load.
Solution Approach 2:
The system performs preliminary image recognition and evaluation before text conversion. By pre-identifying important images using triggering conditions and object recognition, the system prepares a filtered subset of images for real-time processing, reducing overall computation while enabling timely information delivery.
4Productivity
If text descriptions are generated based only on predefined keywords, then the processing is efficient, but key event details may be omitted reducing accuracy
Solution Approach 1:
The patent merges predefined keywords with image recognition results to generate text descriptions. By combining these two information sources, the system maintains processing efficiency through keywords while enhancing accuracy through recognition-based object and event identification, ensuring key details are not omitted.
Solution Approach 2:
The text generation process uses a composite approach combining predefined keywords (for efficiency) with dynamically recognized objects and events (for accuracy). This composite method creates text descriptions that are both efficiently generated and highly accurate, capturing key event details while maintaining productivity.
Data Source
AI summary
The present disclosure provides a method and an electronic apparatus for distributing image description text. The method including: obtaining video data including multiple images; performing image recognition processing on the multiple images to identify images with important information, the important information indicates objects and/or events in the images; in response to the objects and/or events satisfying predetermined triggering conditions, obtaining predefined keywords corresponding to the objects and/or events in the images; generating image description text indicating the objects and/or events based on the obtained predefined keywords and the images with important information; and distributing the generated image description text to terminal devices.