https://github.com/bobld/pdfpigsvmregionclassifier
Proof of concept of a simple SVM Region Classifier using PdfPig and Accord.Net. The objective is to classify each text block in a pdf document page as either title, text, list, table and image.
https://github.com/bobld/pdfpigsvmregionclassifier
accord-net csharp document-layout-analysis machine-learning pdf pdf-document pdfpig publaynet support-vector-machine svm svm-classifier svm-training
Last synced: 8 months ago
JSON representation
Proof of concept of a simple SVM Region Classifier using PdfPig and Accord.Net. The objective is to classify each text block in a pdf document page as either title, text, list, table and image.
- Host: GitHub
- URL: https://github.com/bobld/pdfpigsvmregionclassifier
- Owner: BobLd
- License: mit
- Created: 2019-12-10T10:04:59.000Z (about 6 years ago)
- Default Branch: master
- Last Pushed: 2022-06-23T08:49:08.000Z (over 3 years ago)
- Last Synced: 2025-04-06T07:38:52.356Z (9 months ago)
- Topics: accord-net, csharp, document-layout-analysis, machine-learning, pdf, pdf-document, pdfpig, publaynet, support-vector-machine, svm, svm-classifier, svm-training
- Language: C#
- Homepage:
- Size: 1.13 MB
- Stars: 7
- Watchers: 2
- Forks: 1
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
- License: LICENSE
Awesome Lists containing this project
README
# PdfPig SVM Region Classifier
Proof of concept of a simple Support Vector Machine Region Classifier using [PdfPig](https://github.com/UglyToad/PdfPig) and [Accord.Net](https://github.com/accord-net/framework/). The model was trained on a subset of the [PubLayNet](https://github.com/ibm-aur-nlp/PubLayNet#getting-data) dataset. See their license [here](https://cdla.io/permissive-1-0/).
The objective is to classify each text block using machine learning in a pdf document page as either _title_, _text_, _list_, _table_ and _image_.
The annotions from the dataset (see sample [here](https://github.com/ibm-aur-nlp/PubLayNet/blob/master/examples/samples.json)) were converted to the [PAGE](https://github.com/PRImA-Research-Lab/PAGE-XML) xml format. See the [`PageXmlConverter`](https://github.com/BobLd/PdfPigSvmRegionClassifier/blob/master/PdfPigSvmRegionClassifier/PageXmlConverter.cs) to convert the json file into PAGE xml files. Images from the dataset were not used. You will need to download the pdf documents separately as we leverage the pdf documents features directly instead.
# Labels
Following the [PubLayNet](https://github.com/ibm-aur-nlp/PubLayNet) methodology, the following [categories](https://github.com/ibm-aur-nlp/PubLayNet/tree/master/pre-trained-models) are available:
|Label|id (svm)|
|---:|:---:|
|**title**|0|
|**text**|1|
|**list**|2|
|**table**|3|
|**image**|4|
# Features
## Text
- Character count
- Percentage of numeric characters
- Percentage of alphabetical characters
- Percentage of symbolic characters
- Percentage of bullet characters
- Average delta to average page glyph height
## Paths
- Path count
- Percentage of Bezier curve paths
- Percentage of horizontal paths
- Percentage of vertical paths
- Percentage of oblique paths
## Images
- Image count
- Average area covered by images
## Code
See the [`GenerateData`](https://github.com/BobLd/PdfPigSvmRegionClassifier/blob/master/PdfPigSvmRegionClassifier/GenerateData.cs) class to generate a csv file with the features, using the pdf documents, and their respective PageXml ground truth (one xml document per page). See the [`FeatureHelper`](https://github.com/BobLd/PdfPigSvmRegionClassifier/blob/master/PdfPigSvmRegionClassifier/FeatureHelper.cs) class to easily generate the features vector from a block.
# Results (in sample)
## Accuracy
Model accuracy = 90.898
## Normalised confusion matrix
## Confusion matrix
| |title|text|list|table|image|
|---:|:---:|:---:|:---:|:---:|:---:|
|**title**|9312|1592|19|3|135|
|**text**|1166|37136|988|820|32|
|**list**|0|1|32|0|0|
|**table**|0|16|4|1092|3|
|**image**|0|0|0|0|154|
## Precision, Recall and F1 score
| |Precision|Recall|F1 score|
|---|:---:|:---:|:---:|
|**title**|0.842|0.889|0.865|
|**text**|0.925|0.958|0.941|
|**list**|0.970|0.031|0.059|
|**table**|0.979|0.570|0.721|
|**image**|1.000|0.475|0.644|
## Code
See the [`Trainer`](https://github.com/BobLd/PdfPigSvmRegionClassifier/blob/master/PdfPigSvmRegionClassifier/Trainer.cs) class to **train** and **evaluate** the model.
After training, the SVM model will be saved as a Gzip.
# Usage
Once the training is finished, you can test the classification on a new pdf document by using either [DocstrumBoundingBoxes](https://github.com/UglyToad/PdfPig/blob/master/src/UglyToad.PdfPig/DocumentLayoutAnalysis/DocstrumBoundingBoxes.cs) or [RecursiveXYCut](https://github.com/UglyToad/PdfPig/blob/master/src/UglyToad.PdfPig/DocumentLayoutAnalysis/RecursiveXYCut.cs) to generate the text blocks, and then classify each block.
See [`SvmZoneClassifier`](https://github.com/BobLd/PdfPigSvmRegionClassifier/blob/master/PdfPigSvmRegionClassifier/SvmZoneClassifier.cs) for a demo implementation. The SVM trained model is available [here](https://github.com/BobLd/PdfPigSvmRegionClassifier/tree/master/PdfPigSvmRegionClassifier/model).
# References
- https://visualstudiomagazine.com/articles/2019/02/01/support-vector-machines.aspx
- http://accord-framework.net/docs/html/T_Accord_MachineLearning_Performance_GridSearch_2.htm
- https://github.com/ibm-aur-nlp/PubLayNet