In a recent google ai blog post an transformer based autoencoder is shown that can be used to learn and generate document layouts.
- Using Variational Transformer Networks to Automate Document Layout Design
- Neural Design Network: Graphic Layout Generation with Constraints
- Variational Transformer Networks for Layout Generation Current approaches to layout generation use greedy search algorithms, such as
- beam search
- nucleus sampling
- top-k sampling
- LayoutParser: A Unified Toolkit for Deep Learning Based Document Image Analysis
- Awesome OCR
- Guide to LayoutParser: A Document Image Analysis Python Library
Citation
BibTeX citation:
@online{bochman2021,
author = {Bochman, Oren},
title = {TensorFlow Probability},
date = {2021-06-01},
url = {https://orenbochman.github.io/posts/2021/2021-06-10-layout-models/2021-06-10-layout-models.html},
langid = {en}
}
For attribution, please cite this work as:
Bochman, Oren. 2021. “TensorFlow Probability.” June 1,
2021. https://orenbochman.github.io/posts/2021/2021-06-10-layout-models/2021-06-10-layout-models.html.