whereismyai
deepdoctection
deepdoctection is an open source Python library that orchestrates document layout analysis, OCR, and classification for PDF and scanned files.
deepdoctection is a Python library designed for document understanding and extraction pipelines. It orchestrates layout analysis, table recognition, optical character recognition, and document or token classification across scanned documents and native PDFs. Layout analysis and table extraction run via Detectron2 and Hugging Face Transformers, while OCR integrates with Tesseract, DocTr, or AWS Textract. The library supports models such as LayoutLM, LiLT, BERT, and RoBERTa, and includes tools to fine-tune models and evaluate pipelines. It is built for developers and researchers building document extraction workflows using PyTorch.
deepdoctection at a glance
Key features
- Layout analysis and table recognition using Detectron2 and Transformers
- OCR integration with Tesseract, DocTr, and AWS Textract
- Document and token classification with LayoutLM, LiLT, and BERT-style models
- Native PDF text mining with pdfplumber
- Language detection using xlm-roberta-base-language-detection
- Image deskewing and rotation with jdeskew or Tesseract
- Pipeline evaluation and fine-tuning for detection and classification models
Use it for
- Decomposing PDF documents into structured layout elements
- Extracting tables and text from scanned and digital documents
- Classifying documents and labeling individual tokens
- Fine-tuning object detection and language models on custom document datasets
Alternatives to deepdoctection
- docTR — Extracts text and layout details from documents using deep learning.
- Docling — Provides advanced PDF understanding and local execution for sensitive data.
- PDF Inspector — Detects if a PDF is scanned or text-based to avoid unnecessary OCR costs.
- MinerU — Parses complex documents and tables into structured markdown for agents.
- RapidOCR — It converts PaddleOCR models to ONNX format for fast inference and deployment.
- MarkItDown — Converts documents into LLM-friendly Markdown for better token efficiency.
- MiniCPM OCR — Efficient document parsing with correctness assessment and high throughput.
- OpenDataLoader PDF — It ranks first in benchmarks for layout and table extraction without needing a GPU.
- NuMind — Achieves high accuracy in document extraction and outperforms generalist models.
- Sparrow — Extracts structured data locally from documents without cloud dependencies.
- DocETL — Extracts structured data and links knowledge using large language models.
- LangExtract — It maps every extraction to its exact location in the source text.
All deepdoctection alternatives