Graph Neural Network Worker Task Estimation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing systems struggle to accurately estimate tasks performed by workers without relying on expensive sensors, which can be cumbersome and interfere with the task at hand.
Innovation Solution
A processing system that generates graph data based on the pose of a worker estimated from images, using a neural network including a graph neural network (GNN) to improve task estimation accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If sensors are used to estimate tasks performed by workers, then measurement precision is improved, but device complexity and cost increase
Solution Approach 1:
The patent replaces physical sensors with an image processing system that uses cameras and neural networks to estimate worker tasks. The system captures images of workers and uses pose estimation technology to analyze body movements and infer performed tasks, eliminating the need for mechanical sensors attached to workers.
Solution Approach 2:
The patent creates a virtual model (graph data representation) of the worker's pose based on image data. This virtual model includes nodes representing body parts and edges representing relationships between them, allowing task estimation without physical sensors. The graph data serves as a digital copy that captures essential movement information.
2Measurement precision
If sensors are attached to workers for task estimation, then measurement precision is improved, but ease of operation deteriorates due to interference with task performance
Solution Approach 1:
The patent replaces physical sensors that would be attached to workers with a remote image processing system. Cameras capture images of workers performing tasks, and neural networks analyze these images to estimate tasks without any physical contact or attachment to the worker's body.
Solution Approach 2:
The system creates a digital representation of worker pose and movement through graph data derived from images, rather than using physical sensors on the worker. This virtual model enables accurate task estimation while leaving the worker completely undisturbed during task performance.
Data Source
AI summary
According to one embodiment, a processing system generates first graph data based on a pose of a worker. The pose is estimated based on a first image of the worker. The first graph data includes a plurality of first nodes corresponding respectively to a plurality of joints of the worker, and a plurality of first edges corresponding respectively to a plurality of skeletal parts of the worker. The processing system inputs the first graph data to a neural network including a graph neural network (GNN). The processing system estimates a task being performed by the worker, by using a result output from the neural network.