Addressing posterior collapse by splitting decoders in variational recurrent autoencoders.
Jianyong SunFan SongQiaohong LiPublished in: Neurocomputing (2024)
Keyphrases
- denoising
- image segmentation
- feed forward
- recurrent neural networks
- posterior distribution
- probability distribution
- posterior probability
- probabilistic model
- bayesian framework
- optical flow
- multiscale
- hidden layer
- methods in computer vision
- prior knowledge
- data sets
- variational methods
- feedforward neural networks
- decoding algorithm
- multiplicative noise removal