Tibetan cursive (Betsug)

Free Public AI Model for Handwritten Text Recognition with Transkribus

Tibetan cursive (Betsug)

This model is tailored towards transcribing the handwritten Tibetan cursive script
known as Betsug (dpe tshugs).

This model was created in the framework of the ERC project The Dawn of Tibetan Buddhist Scholasticism (11th-13th c.) (TibSchol) and was released by Rachael Griffiths (rachaelgriffiths1@gmail.com).

This Betsug model uses the Drutsa model (‘Tibetan cursive (Drutsa)’, id: 54525) as a base model.
For best results, apply this model after running the “Tibetan pecha” layout recognition
model (id: 54306).

Ground truth data consists of 93 folios from a selection of 8 Tibetan treatises being explored in the TibSchol project. 85 folios were used in the training set and 8 in the validation set.
Transcripts use the extended Wylie transliteration system in Roman alphabet.
Abbreviations were transcribed as they appear in the manuscripts, for a list of abbreviations see https://github.com/ERC-TibSchol/abbreviations.

The GT also includes 360 images of abbreviations kindly provided by MonlamAI.

If this model is used as base model for your own model, you are kindly requested to mention the model.

The TibSchol project has received funding from the European Research Council (ERC) under the European Union’s Horizon 2020 research and innovation programme (grant agreement No. 101001002).

This model is published within the project team’s responsibility. The European Research Council or the European Commission must not be held responsible for its further use.

Model Overview

Tibetan cursive (Betsug)
TibSchol Project
Model ID:
11th, 13th, 12th
Tibetan cursive (Betsug)
CER on validation set:
3.61 %
Simply upload a picture and test this model

By uploading an image, you accept our terms and conditions and our privacy policy

Tibetan cursive (Betsug) is freely available to everyone

Get started with Transkribus and use it for your own Material
You can use this model to automatically transcribe Handwritten documents with Handwritten Text Recgnition in Transkribus. This model can be used in the Transkribus Expert Client as well as in Transkribus Lite.
This AI model was trained to automatically convert text from images of historical Tibetan cursive (Betsug) documents into editable and searchable text.