Can I read Self-Constructing Graphs for Semantic Labeling on EtoBox?

Self-Constructing Graphs for Semantic Labeling by Mustafa Mohammadi Gharasuie is a document available to read on EtoBox.

What is Self-Constructing Graphs for Semantic Labeling about?

The document proposes a self-constructing graph convolutional network (SCG-Net) for semantic labeling of aerial imagery. SCG-Net uses learnable latent variables to generate embeddings and self-construct an underlying graph directly from input features, without relying on manually built prior knowledge graphs. It transforms image features into a latent graph structure and assigns pixels to graph vertices. Then graph convolutional networks are used to update node features along graph edges. SCG-Net achieves c

Author

Mustafa Mohammadi Gharasuie

Language

EN