State and action space segmentation algorithm in Q-learning.
Akira NotsuHidetomo IchihashiKatsuhiro HondaPublished in: IJCNN (2008)
Keyphrases
- segmentation algorithm
- action space
- state space
- reinforcement learning
- state action
- automatic segmentation
- continuous state spaces
- markov decision processes
- region growing
- optimal policy
- action selection
- reinforcement learning methods
- unsupervised segmentation
- image segmentation
- graph cuts
- segmentation method
- superpixels
- input image
- real valued
- continuous state
- reinforcement learning algorithms
- state and action spaces
- model free
- segmentation scheme
- search space
- dynamic programming
- cooperative
- initial segmentation
- image segmentations
- linear text segmentation
- image segmentation algorithms
- initial state
- stochastic processes
- single agent
- long run
- mobile robot
- pairwise
- search algorithm