A Repo For Document AI

Integrate AI-powered Document Analysis Pipelines

Detectron2 for Document Layout Analysis

🔄 A tool for object detection and image segmentation dataset format conversion.

PubLayNet for huggingface datasets

Diffusion Layout Transformer implementation.

Using a MaskRCNN model trained on the PublayNet dataset with ML.Net in C# / .Net for Document layout analysis and page segmmentation task.

Trained Detectron2 object detection models for document layout analysis based on PubLayNet dataset

Trained Detectron2 object detection models for document layout analysis based on PubLayNet dataset

Extract and convert PubLayNet data to PageXml format

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.

Vision Based Document Layout Detection, Segmentation and context classification using MaskRCNN on Tensorflow-Keras, PyTorch & Detectron2.

ICDAR 2019: MaskRCNN on PubLayNet datasets. Paragraph detection, table detection, figure detection,...

Proof of concept of training a simple Region Classifier using PdfPig and ML.NET (LightGBM). The objective is to classify each text block in a pdf document page as either title, text, list, table and image.