OpenViSTAI · Agents · Blockchain · Crypto
Submit a project
Menu
AI Data & Analytics

Multi Type Td Tsr

Multi Type Td Tsr is an open-source research project that detects tables in document images and recognizes their cell structure, with workflows for deskewing, structure recognition and combined table detection and recognition.

Explore the practical guide ↓

Independent OpenViST project profile, based on linked sources.

Multi Type Td Tsr — OpenViST project profile in AI Data & Analytics.
OpenViST project illustration. See the linked sources for product details.
Understand the project

Practical overview

When it fits

Use it when a document-image workflow needs table and cell boundaries represented in processed images and XML, especially when tables have different border styles.

Who it is for

  • Document-processing teams preparing scanned pages for structured data workflows
  • Computer-vision researchers exploring table detection and structure recognition

Documented uses

  • Deskew document images before table processing
  • Recognize cell layouts in bordered, unbordered and partially bordered tables
  • Run table detection followed by structure recognition and save table and cell bounding boxes
Before you choose

Requirements & limits

The source code is public under the MIT license. The repository links to a Google Colab notebook and documents local Python scripts.

  • The README lists PyTorch 1.7.0, Torchvision 0.8.1, CUDA 10.1 and PyYAML 5.1 as dependencies.
  • The table-detection model is based on Detectron2, with a separate installation guide referenced by the project.
  • Structure-recognition output includes XML bounding boxes for cells; the documented workflow does not specify cell-text extraction.

Check current requirements in the official documentation. Source-based guidance does not mean OpenViST has run or independently tested this project.

Source-based practical guide

Run table structure recognition on a folder of table images

Process table images with a selected structure-recognition method and save processed images and cell bounding boxes.

Before you start

  • Python 3
  • PyTorch 1.7.0, Torchvision 0.8.1, CUDA 10.1 and PyYAML 5.1, as listed by the repository
  • A folder of table images

Steps

  1. Install the documented library dependencies.

  2. Run the command below.

    Documented command
    python3 tsr.py

  3. Set --folder to the input image folder, --type to borderd, unbordered, partially or partially_color_inv, --img_output to the processed-image output folder, and --xml_output to the XML output folder.

Expected outcome

The workflow is documented to produce processed images and XML files containing cell bounding boxes.

Guide limitations

  • The documented output is image processing and bounding-box data; the README does not describe extracting cell text.
  • The dependency list and script instructions do not establish compatibility with other software versions.
Documentation 1 ↗Documentation 2 ↗
Source-based practical guide

Deskew document images before table processing

Apply the repository's image-alignment preprocessing to images in a folder.

Before you start

  • Python 3
  • A folder of document images
  • The repository's listed dependencies: PyTorch 1.7.0, Torchvision 0.8.1, CUDA 10.1 and PyYAML 5.1

Steps

  1. Run the command below.

    Documented command
    python3 deskew.py

  2. Set --folder to the input folder containing document images.

  3. Set --output to the destination folder for deskewed images.

Expected outcome

Deskewed images are written to the specified output folder.

Guide limitations

  • This documented step covers image alignment preprocessing only, not table detection or structure recognition.
Documentation 1 ↗Documentation 2 ↗
Evidence and transparency

Sources & review

Official repository ↗Guide documentation ↗
Source record details
Source availability
Active
Sources checked
17 October 2026
Original website title
Psarpei/Multi-Type-TD-TSR · Official homepage title · retained in its source language
How we review sourcesSuggest a correction