Real-Time Action Tracking via Neural Network and Rule-Based Image Processing
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing image processing technologies are too slow and inefficient to accurately track actions of individuals with respect to objects in real-time, making it difficult to determine relationships between individuals and objects for applications such as frictionless stores and security systems.
Innovation Solution
A system and method for image processing that involves receiving a time-series set of images, identifying actions taken by tracked individuals with respect to tracked items, and using a combination of neural networks and rule-based approaches to determine the relationship between individuals and objects, improving processor throughput and response times.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Speed
If existing image processing technologies are used to identify individuals and objects, then identification can be achieved, but processing speed is too slow and processor throughput is insufficient for real-time applications
Solution Approach 1:
The system segments the complex task of action identification into multiple components: individual detection, object detection, and action classification. By processing these segments separately and in parallel, the system achieves both high speed and accuracy in real-time video analysis
Solution Approach 2:
The patent transitions from analyzing static individual and object detections to analyzing temporal sequences of these detections across multiple video frames. This dimensional transition from spatial to temporal analysis enables action identification while maintaining real-time processing throughput
2Loss of information
If existing image processing technologies are used, then individual and object recognition can be performed, but the system cannot properly discern important actions taken by individuals within the images/video
Solution Approach 1:
The system performs preliminary detection of individuals and objects in each video frame before analyzing their interactions. By pre-identifying these elements and tracking them across frames, the system preserves action information while maintaining high processor throughput for real-time analysis
3Loss of time
If traditional image processing methods are used to determine relationships between individuals and objects, then analysis can be performed, but response time is too slow for real-time applications
Solution Approach 1:
The system continuously tracks individuals and objects across video frames, maintaining persistent identifiers and updating their positions and interactions in real-time. This continuous processing enables fast response times while preserving the accuracy of action tracking through uninterrupted temporal analysis
Data Source
AI summary
Cameras capture time-stamped images of predefined areas. Individuals and item are tracked in the images. A time-series set of images are processed to determine actions taken by the individuals with respect to the items or to determine relationships between the individuals to the items.