Extract text from pdf by dit detection and ocr

The script uses various libraries such as pdf2image, easyocr, ditod and detectron2 for processing.

Detected objects are categorized into "text", "title", "list", "table", and "figure".

The script provides detailed timing information for various processing steps, which can be useful for performance analysis.

Text extraction uses easyocr and the results are further processed using SymSpell for word segmentation and a regular expression for filtering.

1. Installation

pip install -r requirements.txt
python -m pip install 'git+https://github.com/facebookresearch/detectron2.git'
apt-get install poppler-utils

2. Download OCR model

• Please download the weight trained_ocr_cascade_large.pth first.
• Please set the weight path in configs/cascade_dit_large.yaml.

3. Basic usage

python pdf2txt.py --pdf_path path_to_pdf_file  --outputs_dir path_to_output_dir

The output txt file will be stored in path_to_output_dir/txt