AVT: Unsupervised Learning of Transformation Equivariant Representations by Autoencoding Variational Transformations.
Guo-Jun QiLiheng ZhangChang Wen ChenQi TianPublished in: CoRR (2019)
Keyphrases
- unsupervised learning
- transformation rules
- invariant representations
- geometric transformations
- intermediate representations
- deep belief networks
- image segmentation
- image transformations
- linear transformation
- object recognition
- semi supervised
- model selection
- expectation maximization
- dimensionality reduction
- free energy
- rotation invariant
- multiscale
- optical flow
- optical flow computation
- feature selection
- computer vision
- methods in computer vision
- data sets