Video Event Graph for Accurate Recognition
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current video recognition technologies primarily focus on action or scene recognition, lacking an effective method for video event recognition that can accurately and efficiently understand deep video content.
Innovation Solution
A video event recognition method that constructs a video event graph using a knowledge graph, incorporating argument roles and arguments, and employs graph embedding with Convolutional Neural Networks (CNNs) to select recognized events from the graph based on acquired arguments.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If video action recognition or video scene recognition methods are used, then video recognition can be performed, but accurate video event recognition cannot be achieved
Solution Approach 1:
The patent segments video event recognition into multiple argument roles (e.g., agent, instrument, patient, location, time) and recognizes each role separately through dedicated neural network branches, then integrates them to achieve comprehensive event recognition with high accuracy
Solution Approach 2:
The patent transitions from traditional 2D video frame analysis to graph-based event structure representation, adding dimensional depth by modeling semantic relationships between event arguments, enabling more accurate event understanding
2Measurement precision
If complex video event recognition methods are developed, then event recognition accuracy improves, but computational complexity increases
Solution Approach 1:
The patent merges multiple neural network branches that process different argument roles into a unified graph-based event structure, integrating their outputs to achieve accurate event recognition while maintaining systematic organization
Solution Approach 2:
The patent designs a universal graph-based event representation framework that can handle various types of video events through a common structure, enabling the system to process different event types with a single unified approach
Data Source
Figure 1
Figure 2(a)~3
Figure 4
AI summary
The present disclosure discloses a video event recognition method and apparatus, an electronic device and a storage medium, and relates to the fields of knowledge graphs, deep learning and computer vision. The method may include: constructing a video event graph, each event in the video event graph including: M argument roles of the event and respective arguments of the argument roles, M being a positive integer greater than one; acquiring, for a to-be-recognized video, respective arguments of the M argument roles of a to-be-recognized event corresponding to the video; and selecting, according to the arguments acquired, an event from the video event graph as a recognized event corresponding to the video. Accurate and efficient video event recognition can be implemented by using the solution of the present disclosure.