Partial Policy-based Reinforcement Learning for Anatomical Landmark Localization in 3D Medical Images.
Walid Abdullah AlIl Dong YunPublished in: CoRR (2018)
Keyphrases
- landmark localization
- reinforcement learning
- optimal policy
- policy search
- action selection
- markov decision process
- face detection
- partially observable environments
- actor critic
- policy gradient
- function approximation
- reinforcement learning problems
- control policy
- function approximators
- markov decision problems
- reward function
- approximate dynamic programming
- control policies
- policy evaluation
- three dimensional
- markov decision processes
- medical images
- partially observable
- state action
- state space
- policy iteration
- action space
- reinforcement learning algorithms
- temporal difference
- state dependent
- multi agent
- learning process
- partially observable domains
- state and action spaces
- model free
- continuous state
- agent learns
- dynamic programming
- long run
- exploration exploitation tradeoff
- learning algorithm
- rl algorithms
- partially observable markov decision processes
- infinite horizon
- finite state
- anatomical structures
- machine learning
- continuous state spaces
- control problems
- optimal control
- mr images