Misalignment Recognition Using Markov Random Fields With Fully Connected Latent Variables for Detecting Localization Failures.
Naoki AkaiLuis Yoichi MoralesTakatsugu HirayamaHiroshi MurasePublished in: IEEE Robotics Autom. Lett. (2019)
Keyphrases
- latent variables
- fully connected
- probabilistic model
- latent variable models
- random variables
- topic models
- prior knowledge
- scale free
- hidden variables
- conditional random fields
- approximate inference
- feature extraction
- structural svms
- hidden layer
- observed variables
- linear combination
- generative model
- bayesian networks
- training data
- image segmentation