Document Layout Analysis
36 papers with code • 4 benchmarks • 9 datasets
"Document Layout Analysis is performed to determine physical structure of a document, that is, to determine document components. These document components can consist of single connected components-regions [...] of pixels that are adjacent to form single regions [...] , or group of text lines. A text line is a group of characters, symbols, and words that are adjacent, “relatively close” to each other and through which a straight line can be drawn (usually with horizontal or vertical orientation)." L. O'Gorman, "The document spectrum for page layout analysis," in IEEE Transactions on Pattern Analysis and Machine Intelligence, vol. 15, no. 11, pp. 1162-1173, Nov. 1993.
Image credit: PubLayNet: largest dataset ever for document layout analysis
Benchmarks
These leaderboards are used to track progress in Document Layout Analysis
| Trend | Dataset | Best Model | Paper | Code | Compare |
|---|---|---|---|---|---|
|
|||||
|
|||||
|
|||||
|
Libraries
Use these libraries to find Document Layout Analysis models and implementations
PubLayNet
RVL-CDIP
HJDataset
ICDAR 2021
UrduDoc
Most implemented papers
Training data-efficient image transformers & distillation through attention
In this work, we produce a competitive convolution-free transformer by training on Imagenet only.
LayoutLM: Pre-training of Text and Layout for Document Image Understanding
microsoft/unilm
•
•
In this paper, we propose the \textbf{LayoutLM} to jointly model interactions between text and layout information across scanned document images, which is beneficial for a great number of real-world document image understanding tasks such as information extraction from scanned documents.
BEiT: BERT Pre-Training of Image Transformers
We first "tokenize" the original image into visual tokens.
PubLayNet: largest dataset ever for document layout analysis
Deep neural networks that are developed for computer vision have been proven to be an effective method to analyze layout of document images.
LayoutLMv2: Multi-modal Pre-training for Visually-Rich Document Understanding
Pre-training of text and layout has proved effective in a variety of visually-rich document understanding tasks due to its effective model architecture and the advantage of large-scale unlabeled scanned/digital-born documents.
dhSegment: A generic deep-learning approach for document segmentation
In recent years there have been multiple successful attempts tackling document processing problems separately by designing task specific hand-tuned strategies.
Combining Visual and Textual Features for Semantic Segmentation of Historical Newspapers
The massive amounts of digitized historical documents acquired over the last decades naturally lend themselves to automatic processing and exploration.
A Large Dataset of Historical Japanese Documents with Complex Layouts
dell-research-harvard/HJDataset
•
•
Deep learning-based approaches for automatic document layout analysis and content extraction have the potential to unlock rich information trapped in historical documents on a large scale.
CDeC-Net: Composite Deformable Cascade Network for Table Detection in Document Images
mdv3101/CDeCNet
•
•
Localizing page elements/objects such as tables, figures, equations, etc.
DiT: Self-supervised Pre-training for Document Image Transformer
microsoft/unilm
•
•
We leverage DiT as the backbone network in a variety of vision-based Document AI tasks, including document image classification, document layout analysis, table detection as well as text detection for OCR.
