Dataset

Bharat-Scene-Text-Dataset

Name: Bharat-Scene-Text-Dataset
Creator: Bhashini-IITJ

by Bhashini-IITJ

Bharat Scene Text Dataset (BSTD) v17. 58 BSTD is a benchmark for Indian-language scene text: 6,582 scene images from Wikimedia Commons with 126,292 manually annotated text regions (polygons, transcriptions, and script labels).

No licenseobject-detection 23 0100K < n < 1M rows

published 12 Oct 2026

Open on Hugging Face

Preview

The dataset does not exist, or is not accessible without authentication (private or gated). Please check the spelling of the dataset name or retry with authentication.

Cite this

No licenseimagetextCite

Tags

task_categories:object-detectiontask_categories:image-to-texttask_categories:image-classificationmultilinguality:multilinguallanguage:aslanguage:bnlanguage:enlanguage:gulanguage:hilanguage:knlanguage:mllanguage:mrlanguage:orlanguage:palanguage:talanguage:telicense:othersize_categories:100K<n<1Mformat:parquetmodality:imagemodality:textlibrary:datasetslibrary:pandaslibrary:polarslibrary:mlcroissantregion:usocrscene-textscene-text-detectionscene-text-recognitionscript-identificationindicwikimedia-commons