PocketFinderGNN: A manufacturing feature recognition software based on Graph Neural Networks (GNNs) using PyTorch Geometric and NetworkX.
Igor BetkierMateusz OszczypalaJanusz PobozniakSergiusz SobieskiPrzemyslaw BetkierPublished in: SoftwareX (2023)
Keyphrases
- neural network
- pattern recognition
- preprocessing stage
- object recognition
- recognition rate
- recognition accuracy
- model based recognition
- graph theory
- image features
- software development
- recognition process
- vector representation
- software tools
- feature vectors
- fuzzy logic
- computer systems
- feature extraction
- directed graph
- graph structure
- point distribution
- discrimination power
- source code
- graph representation
- software systems
- action recognition
- bipartite graph
- graph mining
- graph model
- spanning tree
- graph theoretical
- artificial neural networks
- software projects
- neural network model
- highly discriminative
- activity recognition
- consistency check
- structured data
- extended gaussian image
- graph matching
- feature analysis
- graph databases
- directed acyclic graph
- training process
- manufacturing systems
- connected components
- software architecture
- self organizing maps
- random walk
- computer vision