GNOSIS: Proactive Image Placement Using Graph Neural Networks & Deep Reinforcement Learning.
Theodoros TheodoropoulosAntonios MakrisEvangelos PsomakelisEmanuele CarliniMatteo MordacchiniPatrizio DazziKonstantinos TserpesPublished in: CLOUD (2023)
Keyphrases
- neural network
- reinforcement learning
- image data
- input image
- image collections
- single image
- image classification
- image features
- inexact graph matching
- image content
- image segmentation
- image pixels
- hierarchical data structure
- image retrieval
- image analysis
- high resolution
- low level
- test images
- image matching
- multiscale
- template matching
- similarity measure
- function approximation
- adjacency graph
- image representation
- random walk
- markov random field
- state space
- genetic algorithm
- learning algorithm
- computer vision
- model free
- image processing
- graph theory
- multi agent
- pattern recognition
- directed graph
- optimal policy
- region of interest
- markov decision processes
- segmentation algorithm
- spatial information
- hough transform
- segmentation method