Social Media Event Detection via Visual Semantic Extraction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing approaches for event detection in social media rely solely on text-based inputs and lack the ability to quantify event characteristics from images, limiting their information and being language-specific, thus failing to effectively capture and analyze visual data for event detection.
Innovation Solution
A computer-implemented method that extracts visual semantic concepts from social media images, differentiates event semantic concepts from background signals, and retrieves relevant images to present a visual description of detected events, utilizing techniques like support vector machines and machine learning algorithms to analyze image semantics over time.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Loss of information
If text-based input with SVM classifier is used for event detection, then language-specific event detection can be performed, but the system cannot quantify event characteristics from images and loses visual information
Solution Approach 1:
The patent combines text-based event detection with image-based visual concept detection into a unified system. The event detection module processes text inputs while the visual concept detection module processes images, and both results are integrated to provide comprehensive event characterization, thereby recovering visual information that was previously lost
Solution Approach 2:
The system is designed to handle multiple types of inputs (text and images) and perform multiple functions (event detection and visual concept extraction) through a unified architecture. This multi-functional approach allows the system to process diverse social media content types without requiring separate specialized systems
2Loss of information
If only text-based data is processed, then language-specific event detection is achieved, but the system lacks comprehensive understanding of event characteristics
Solution Approach 1:
The system segments the event detection process into independent modules: text-based event detection, image-based visual concept detection, and result integration. This segmentation allows parallel processing of different data types, reducing overall processing time while maintaining comprehensive event characteristic analysis
3Measurement precision
If visual semantic concepts are extracted and differentiated from background signals, then event detection accuracy improves, but system complexity increases
Solution Approach 1:
The system employs feedback mechanisms where detected visual concepts and event signals are continuously refined through comparison with background signals. The differentiation process uses feedback from signal analysis to adjust detection thresholds and improve accuracy while managing complexity through iterative optimization
Data Source
AI summary
Techniques for detecting an event via social media content. A method includes obtaining multiple images from at least one social media source; extracting at least one visual semantic concept from the multiple images; differentiating an event semantic concept signal from a background semantic concept signal to detect an event in the multiple images; retrieving one or more images associated with the event semantic concept signal; grouping the one or more images associated with the event semantic concept signal; annotating the group of one or more images with user feedback; and displaying the annotated group of one or more images as a visual description of the detected event.