Scalable Font Reconstruction with Dual Latent Manifolds.
Nikita SrivatsanSi WuJonathan T. BarronTaylor Berg-KirkpatrickPublished in: EMNLP (1) (2021)
Keyphrases
- smooth manifold
- euclidean space
- image reconstruction
- reconstruction process
- manifold learning
- highly scalable
- three dimensional
- discrete tomography
- character recognition
- latent variables
- high dimensional
- reconstruction method
- arbitrary dimension
- geometric structure
- data sets
- feature space
- web scale
- compressed sensing
- latent space
- low dimensional
- lie group
- linear programming
- input image
- document image understanding
- multiresolution