IIEST-Indic-HW is a multilingual handwritten Indic dataset. This dataset consists of samples from 170 writers written in Bengali and Hindi.
Each document in the dataset has a Document ID and each page in the document has a page ID. A document consists of 10 pages with the first 5 pages containing text written in Bengali and the next 5 pages containing text written in Hindi.
A sample filled page in Bengali and Hindi is shown in Figure 1.a and Figure 1.b, respectively.
138 writers out of 170 have copied both the Bengali and Hindi text and contributed at least 9 pages. So each of the 138 writers has contributed either 9 or 10 pages. 16 writers have copied 5 pages of the Bengali part of the document, and the remaining 16 writers have copied 5 pages of the Hindi part of the document.
Each page of the document was scanned with a resolution of 300 dpi and a colour depth of 24 bit. The images are stored as LZW compressed TIFFs. Each page is labelled as (Writer ID)-(Page ID)_(optional part number)_(B/H). The Writer ID is the same as the document ID, since each document is assigned to a particular writer. The Page ID denotes the ID of each page in the document. The part number is optional and is only included when a writer required more than the space provided to complete the content of a single page. The letters B and H denote whether the content of the page is in Bengali or Hindi, respectively.
We have prepared a lite version of the IIEST-Indic-HW dataset consisting of samples from 20 writers.
To obtain the entire dataset we request you to drop an email to one of the authors below.
We kindly request anyone using this dataset to include a reference to the following paper -
Related Publications:
Namrata Das, Samit Biswas, "Writer Identification from Multilingual Handwritten Indic Text Document Images ", ICVGIP 2025 (Accepted for Publication)
Namrata Das and Samit Biswas. 2025. Writer Identification from Multilingual Handwritten Indic Text Document Images. In Indian Conference on Computer Vision, Graphics, and Image Processing (ICVGIP 2025), December 17–20, 2025, Mandi, India. ACM, New York, NY, USA, 9 pages. https://doi.org/10.1145/3774521.3774586
Contact:
Namrata Das
Email: namratadas628@gmail.com
Samit Biswas
Email: samit@cs.iiests.ac.in
