End-to-end semantic face segmentation with conditional random fields as convolutional, recurrent and adversarial networks.
Umut GüçlüYagmur GüçlütürkMeysam MadadiSergio EscaleraXavier BaróJordi GonzàlezRob van LierMarcel A. J. van GervenPublished in: CoRR (2017)
Keyphrases
- end to end
- conditional random fields
- image labeling
- high bandwidth
- segmentation method
- crf model
- wireless ad hoc networks
- internet protocol
- restricted boltzmann machine
- sequence labeling
- information extraction
- hidden markov models
- probabilistic model
- graphical models
- random fields
- structured prediction
- segmentation algorithm
- generative model
- markov random field
- image segmentation
- level set
- higher order
- ad hoc networks
- admission control
- approximate inference
- pairwise
- differentiated services
- semi markov
- natural language
- energy function
- superpixels
- multiscale
- structured learning
- congestion control
- network structure
- computer networks
- object segmentation
- web page prediction
- loopy belief propagation
- image processing
- transport layer