Find research datasets worth reusing
Search datasets from major research repositories and use ShareScore to quickly assess how well each record supports discovery, access, and reuse.
56
datasets available to search
ShareScore release 0.9.0
Dataset results
56 results for “Handwriting”
Handwriting of the papal chancery and beyond: Dataset
<p><strong>Handwriting of the papal chancery and beyond: Dataset</strong></p> <p>D. Stutzmann</p> <p>The original data was originally produced in Oct.-Nov. 2019 for the paper "Les écritures de la chancellerie pontificale dans le paysage européen (XIIe–XVe siècles)" at the conference "Les actes pontificaux. Un trésor à exploiter" held at the German Historical Institute in Paris on Nov. 28th, 2019.</p> <p>The text of the paper is published on HAL: <a href="https://hal.archives-ouvertes.fr/halshs-03628606">https://hal.archives-ouvertes.fr/halshs-03628606</a>. Parts of the present corpus (preliminary code and partial data) were released on Github on Jan. 4th, 2022 (<a href="https://github.com/oriflamms/RegVat_ArchNatJJ_MOM_LBA_2019">https://github.com/oriflamms/RegVat_ArchNatJJ_MOM_LBA_2019</a>).</p> <p>The study has been extended and deepened for a more comprehensive publication as an article in the proceedings.</p> <p>The analyzed corpora encompass images from</p> <ul> <li>Registra Vaticana (Vaticano, Archivio Apostolico, Reg. Vat.)</li> <li>Registers of French royal chancery (Paris, Archives Nationales, JJ series)</li> <li>images from Monasterium.net and Lichtbildarchiv</li> </ul> <p>The present dataset does not contain the images.</p> <p>The present dataset contains</p> <ol> <li>folder /_initial_corpus/: metadata on the "original corpus" that I intended to use</li> <li>folder /data/: (a) the original output data as produced by the Computer Vision library processing the images; (b) the data with its metadata.</li> <li>folder /rds/: RDS files produced as part of the analysis with the R software (cf. https://www.r-project.org/)</li> <li>folder /images/: figures and illustrations of the article</li> <li>folder /plotly/: corresponding interactive visualisations of the figures</li> </ol> <p>At the root, along witht this README file, the R code to produce the statistics, RDS files and images.</p> <p>For this library, see Nicolaou, Anguelos, A. D. Bagdanov, Marcus Liwicki, et D. Karatzas. « Sparse radial sampling lbp for writer identification ». In 2015 13th International Conference on Document Analysis and Recognition (ICDAR), 716-20, 2015., cf. <a href="https://doi.org/10.48550/arXiv.1504.06133">https://doi.org/10.48550/arXiv.1504.06133</a></p>
MPS Data set with images of medieval charters for handwriting-style based dating of manuscripts
<pre>The MPS benchmark data set for handwritten manuscript dating ____________________________________________________________ This data set is collected for the Dutch NWO project: Medieval Paleographical Scale (MPS) by Petros Samara Project website: http://application02.target.rug.nl/monk/Projects/MPS/ Copyright (c) Huygensinstituut, Den Haag, 2016 University of Groningen, 2016. All rights reserved. Organisation of the data: Each .tar.gz file contains a number of NetPBM images. The format is chosen because of its simplicity. Also, there is no doubt about lossy compression in the processing chain. The file names are of the format 'MPS<year>_<seqnr>.ppm', for example, 'MPS1300_0056.ppm'. Note: the files are not in a separate directory, they will be extracted in place. However, due to the unique naming, there is no problem extracting them in one single current (destination) directory. The actual type of the image can be gray scale (.pgm) or color (.ppm), in '8-bit DirectClass' according to ImageMagick's 'identify' tool. The images were cropped out of larger photographs because of irrelevant elements such as a Kodak color calibrator and non-text content such as supporting surface (table) backgrounds, seals (emblems), ribbons, etc. No effort has been made to obtain a balanced set of samples over years: the given frequencies of occurrence in archives are used. There is evidently less data in years before 1375 A.D. while some periods provides us with ample data for historical reasons (e.g, 1450 A.D.). It would have been a pity if the scarce years had determined and limited the size of this data set. Selection criteria for data reduction, whether random or systematic, would have been arbitrary. In any case, these images were used in our publications, such that the performance results of future attempts on manuscript dating can be compared with earlier results. The performances that have been reached using our algorithms are in the order of an MAE (mean average error) of 10 years. If you have any questions, please contact us: Sheng He (heshengxgd@gmail.com) Petros Samara (petros.samara@huygens.knaw.nl) Jan Burgers (jan.burgers@huygens.knaw.nl) Lambert Schomaker (L.Schomaker@ai.rug.nl) Please cite our papers if you use this data set: [1] Sheng He, Petros Samara, Jan Burgers, Lambert Schomaker. Image-based historical manuscript dating using contour and stroke fragments. Pattern Recognition(PR), Vol. 59, pp. 159-171, 2016 [2] Sheng He, Petros Samara, Jan Burgers, Lambert Schomaker. Towards style-based dating of historical documents. International Conference on Frontiers in Handwriting Recognition(ICFHR), Crete, Greece, 2014 [3] Sheng He, Petros Samara, Jan Burgers, Lambert Schomaker. Multiple-Label Guided Clustering Algorithm for Historical Document Dating and Localization IEEE Trans. on Image Processing, Vol. 25(11), Nov. 2016. http://ieeexplore.ieee.org/document/7551181/</pre> <p>Data are collected thanks to Dutch NWO grant project 380-50-006</p>
ImUnipen image data set for writer identification (N=208) - vectorial handwriting converted to usable images
<p><br> ==============<br> Terms of Usage<br> ==============</p> <p>The ImUnipen data set is intended for non-commercial, scientific use,<br> and is distributed under auspices of the Unipen Foundation.</p> <p>Please always refer to the following paper in IEEE PAMI when using<br> the ImUnipen data set:</p> <p> Bulacu, M.; Schomaker, L.<br> Text-Independent Writer Identification and Verification<br> Using Textural and Allographic Features<br> Pattern Analysis and Machine Intelligence, IEEE Transactions on<br> Volume 29, Issue 4, April 2007 Page(s):701 - 717</p> <p>The ImUnipen data set is derived from the Unipen (unipen.org)<br> data set of on-line (i.e., vectorial, xy) handwriting.<br> The xy-coordinates and a line-generator algorithm are used<br> to generate a raster image, as if the data were optically scanned.</p> <p>Contents: for 208 writers, there are two PNG images per writer of<br> an artificially constructed table of naturally written words (49MByte).<br> These words are pasted onto a white page. For systematics reasons,<br> we call such a page a Paragraph, see below.</p> <p>The file names are organized as (example):</p> <p> Writ990221.Doc01.Par00.png<br> Writ990221.Doc01.Par01.png</p> <p> meaning: writer number 990221, document 01 (there exists only Doc01)<br> and the image with artificial "paragraph" of isolated words "Par00"<br> and "Par01".</p> <p>The Par00 and Pa01 images are typically used as the query<br> and best match in a leave-one-out setting for writer identification.<br> For instance, Par00 is the query, and Par01 is added to the total set<br> of all other images as the attractor for an identification search.</p> <p>For these experiments, word labels are not given in this data set,<br> on purpose, as the goal is to test recognition-free writer identification<br> methods.</p> <p>For a description of the regular<br> Unipen data set, please visit http://unipen.org</p> <p>Lambert Schomaker constructed this set in 2005</p>
Portuguese Handwriting 16th-19th c.
<p>All data were imported from the platform <a href="transkribus.org" target="_blank" rel="noopener">Transkribus</a> on which the AI model for automatic transcription “Portuguese Handwriting 16<sup>th</sup>-19<sup>th</sup> c.” was last trained in July 2023 with the recognition engine Pylaia, and can now be used.</p> <p>The data are divided into ten folders, according to the total number of the trainings, from the initial to the definitive one, plus one set for final validation. The eight previous trainings were realized between June 2022 and May 2023. The history of all trainings can be read on <a href="https://traprinq.hypotheses.org/" target="_blank" rel="noopener">e-Inquisition</a>. Each of these folders corresponds to one collection in the platform; every collection has a number of documents; every document has a number of images, or pages, as indicated below.</p> <p>The ten uploaded folders (zip) are distributed as follows:</p> <p>—nine Training Sets (TS) (ca 92% of the whole data; status of the transcriptions from the TS: Ground Truth);</p> <p>—the final Validation Set (VS) (ca 8% of the whole data; status of the transcriptions from the VS: Ground Truth).</p> <p>All TS folders contain only the new data added to the following training (thus added to the previous data).</p> <p>Only the last VS, which is complete (505 p.), is provided.</p> <p>One document = images / transcribed pages (Ground Truth: transcription made by the members of TraPrInq project (Transcrever os processos da Inquisição portuguesa, 1536-1821 | Transcribing the court records of the Portuguese Inquisition, 1536-1821), which lasted from January 2023 to July 2024.</p> <p>The majority of the documents are titled as follows: IL_number = document extracted from a trial record (<em>processo</em>) by the Inquisition of Lisbon_number of the <em>processo</em>; other titles: IC_ = Inquisition of Coimbra; IE_ = Inquisition of Évora.</p> <p>Total of transcribed pages: 6,417.</p> <p>The quality of the images in the data (jpg) is equal to that of the images used for automatic transcription.</p> <p>All digitized images can be found on the <a href="https://digitarq.arquivos.pt/" target="_blank" rel="noopener">catalog of the Portuguese National Archives</a> (Arquivo Nacional da Torre do Tombo, ANTT).</p> <p>Available data (10 zip files, total size 6.7 GB):</p> <p>Training Set1: 698 pages/images</p> <p>Training Set2: 984 pages/images</p> <p>Training Set3: 869 pages/images</p> <p>Training Set4: 926 pages/images</p> <p>Training Set5: 631 pages/images</p> <p>Training Set6: 665 pages/images</p> <p>Training Set7: 564 pages/images</p> <p>Training Set8: 549 pages/images</p> <p>Training Set9: 531 pages/images</p> <p>Validation Set_Final: 505 pages/images</p> <p>2-one pdf file:</p> <p>Paleographical criteria used by the team for the transcription of the documents; list of characters (in Portuguese).</p>
The e-NDP project : collaborative digital edition of the Chapter registers of Notre-Dame of Paris (1326-1504). Ground-truth for handwriting text recognition (HTR) on late medieval manuscripts.
<p>The <a href="https://endp.hypotheses.org/">e-NDP project</a>, funded by the ANR, is led by the <a href="https://lamop.hypotheses.org/6870">LaMOP</a> (Julie Claustre and Darwin Smith).</p> <p>The project's partners are the Archives nationales, the Bibliothèque nationale de France (Department of Manuscripts, Bibliothèque de l'Arsenal), the École nationale des chartes and the Bibliothèque Mazarine.</p> <p>The e-NDP project aims at renewing our knowledge on <strong>Notre-Dame de Paris cathedral</strong> through the creation of a collaborative digital edition of the registers of its Chapter (1326-1504, <em>AN LL 105-128</em>), the community of 51 canons meeting three times a week on set days to take all administrative, financial and practical decisions pertaining to the cathedral, its estate and the society living in its cloister. This corpus has never been the object of a comprehensive study to understand the workings and history of this urban enclave and powerful community. The collaborative digital edition is based on a process of<strong> handwriting text recognition (HTR)</strong>, tested and supervised by scholars, researchers and engineers combining expertise in Medieval history, paleography, philology and digital humanities. The edition shall allow a better insight into the Chapter’s administration, into its economical and political power within Paris, and the relationships it maintained with other institutions in the city.</p> <p> </p> <p><strong>Section 1 : The e-NDP ground-truth dataset for Handwriting text recognition.</strong></p> <p>The full e-NDP corpus kept today in the French National Archives and was entirely digitized and described in its <a href="https://www.siv.archives-nationales.culture.gouv.fr/siv/rechercheconsultation/consultation/ir/consultationIR.action?formCaller=GENERALISTE&irId=FRAN_IR_059635">catalog</a> in 2022.</p> <p>The first major goal of the e-NDP projet is to propose a first automatic transcription of the 14k pages composing the 26 chapter registers. To achieve this goal representative samples from each one of the volumes were selected and transcribed in order to train a specialized HTR model able to propose a high quality automatic transcription. The collected ground-truth released on this repository currently has <strong>512 pages from the 26 registers</strong> of the cathedral chapter preserved in the National Archives (LL105 - LL128, <strong>1326-1504</strong>). The transcriptions were manually completed in <strong>two rounds</strong> by a group of 12 contributors, historians and paleographers, over the course of 2021-2022 using <a href="https://escriptorium.paris.inria.fr/">eScriptorium </a>as annotation environment. </p> <p> </p> <p><strong>Ground-truth features :</strong></p> <p><br> <em>Number of hands </em>: according to our estimates no fewer than 18 main hands were involved in the writing of the registers during the medieval period. </p> <p><em>Language</em> : More than 98% of the content of the registers was written in Latin, the rest in French. The exact percentage is hard to estimate because the vernacular language is often used in formulae, notes and comments. It is rare to find entire pages or blocks written in French. </p> <p><em>Script family</em> : The registers were written using a Cursive script (ca. late XIIIe - XVIe).</p> <p><em>Documental typology</em> : The volumes containing the chapter conclusions were conceived to serve as memorial records, but above all as documents for regular use and consultation in the daily practice of administration and management. In diplomatics the notion of "documentary manuscripts" is used to describe this kind of sources also by opposition to books and litterary or normative manuscripts.</p> <table align="center"> <caption><strong>Ground truth statistics</strong></caption> <tbody> <tr> <th>Text units</th> <th>Count</th> </tr> <tr> <td>Pages</td> <td>512</td> </tr> <tr> <td>Annotated regions (see section 2)</td> <td>2448</td> </tr> <tr> <td>Lines of text</td> <td>34231</td> </tr> <tr> <td>Tokens</td> <td>205083</td> </tr> <tr> <td>Characters</td> <td>3320407</td> </tr> </tbody> </table> <p> </p> <p><strong>Rules of transcription :</strong></p> <ul> <li>The abbreviations have been resolved, both those by suspension (<code>facimꝰ</code> ---> <code>facimus</code>) and by contraction (<code>dñi</code> --> <code>domini</code>). Likewise, those using conventional signs (<code>⁊</code> --> <code>et</code> ; <code>ꝓ</code> --> <code>pro</code>) have been resolved. </li> <li>The named entities (names of persons, places and institutions) have been <code>capitalized</code>. The beginning of a block of text as well as the original capitals used by the notary are also capitalized.</li> <li>The consonantal <code>i</code> and <code>u</code> characters have been transcribed as <code>j</code> and <code>v</code> in both French and Latin.</li> <li>The punctuation marks used in the text: <code>.</code> and <code>/</code> have been transcribed, but the transcription has not been standardized with modern punctuation.</li> <li>Corrections and words that appear cancelled in the manuscript have been transcribed surrounded by the sign <code>$</code> at the beginning and at the end.</li> <li>More specific transcription rules can be found into the file <code>transcription_guidelines.pdf</code></li> </ul> <p> </p> <p><strong>Section 2. e-NDP Layout Segmentation.</strong></p> <p>Layout segmentation is a compulsory step before HTR recognition in order to distinguish sections and regions inside a document. This process intend to separate interdependant page zones to produce a recognition in a section-sequence order and not in a line-sequence order which mix textual and peri-textual content.</p> <p>The regions of 364 pages (see <code>GT-layout_list</code>) of the e-NDP corpus were annotated using a 5 sections vocabulary (see <code>endp_layout_regions</code>) in order to describe the page distribution in all the 26 volumes :</p> <ol> <li><em>Block</em> : All the central text blocks, that normally corresponds to the main content called "conclusions" in registers.</li> <li><em>Liste</em> : List of names of the canons who were present during the meeting. Normally located before the <em>conclusions</em>.</li> <li><em>Entrée</em> : Marginal notes or entries to inform about the content of <em>conclusions</em>.</li> <li><em>Date</em> : Paragraph contending the date. Normally at the head of a <em>conclusion</em>, but separate of the main body.</li> <li><em>Numérotation</em> : Page numbers in roman or arabic. Usually appear in the top corners of the pages.</li> </ol> <table align="center"> <caption><strong>Layout GT statistics</strong></caption> <tbody> <tr> <th>Region</th> <th>Count</th> </tr> <tr> <td>block</td> <td>833</td> </tr> <tr> <td>liste</td> <td>431</td> </tr> <tr> <td>date</td> <td>448</td> </tr> <tr> <td>entrée</td> <td>205</td> </tr> <tr> <td>numérotation</td> <td>531</td> </tr> </tbody> </table> <p> </p> <p><strong>Section 3. The e-NDP HTR modeling.</strong></p> <p>The e-NDP project has progressively trained several HTR models adapted to work on late medieval cursive in order to accelerate the production of ground truth. Currently the best model delivers an average <strong>CER (Character error ratio) of 9.7%</strong> in handwriting recognition on the 26 registers (see <code>endp_learning_curve</code>) and can serve as generalist model for other manuscripts of the same period and similar script family. These models and their training implementation details can be found in the project's github <a href="https://github.com/chartes/e-NDP_HTR">repository</a>. </p> <p>Additionally, the automatic HTR transcriptions of the 26 registers (14k pages, 4.5M tokens) enriched with lexical and semantical information has been the subject of a first <a href="https://nosketch-engine.lamop.fr/#dashboard?corpname=endp">online publication</a> using the NoSketch engine that allows advanced data mining based on the combination of data, metadata and NLP features. </p> <p> </p> <p><strong>Section 4. Dataset content.</strong></p> <p>This zip dataset contains :</p> <p>- <code>HTR_ground_truth</code> : Two folders containing the jpg / jpeg images and their curated transcriptions in PAGE XML format.</p> <p>- <code>images_docs</code> : 4 files illustrating the different phases of the project (list of GT for layout segmentation, layout ontologie, transcription guideline and HTR evaluation curves)</p>
BanglaWriting Words Dataset: A Collection of Isolated Word Images from the BanglaWriting multi-purpose Bangla offline-handwriting dataset (WoBW)
<p>The WoBW (Words from BanglaWriting) dataset is a curated collection of isolated word images, adapted from the original BanglaWriting dataset (url: https://data.mendeley.com/datasets/r43wkvdk4w/1).</p> <p>WoBW focuses on individual words extracted from handwritten Bangla text samples in the BanglaWriting corpus, making it a valuable resource for research in word-level Bangla handwriting recognition and related natural language processing tasks.</p> <p>Mridha, Dr. M. F.; Quwsar Ohi, Abu; Ali, M. Ameer; Emon, Mazedul Islam; Kabir, Md Mohsin (2020), “BanglaWriting: A multi-purpose offline Bangla handwriting dataset”, Mendeley Data, V1, doi: 10.17632/r43wkvdk4w.1</p>
Wikipedia rendered as synthetic handwriting
<p>This is the synthetic handwriting data used to pre-train <strong>Dessurt</strong> (<a href="https://arxiv.org/abs/2203.16618">https://arxiv.org/abs/2203.16618</a>).</p> <p>It is text sampled from Wikipedia and generated with the method described in "Text and Style Conditioned GAN for Generation of Offline Handwriting Lines" (<a href="https://arxiv.org/abs/2009.00678">https://arxiv.org/abs/2009.00678</a>). More data can be quite easily obtained using this code: <a href="https://github.com/herobd/handwriting_line_generation">https://github.com/herobd/handwriting_line_generation</a></p> <p>Inside the tar is a single directory with ~800k generated handwriting line images ("sample_0.png", "sample_123.png", "sample_3292524.png", etc.), and "OUT.txt" which has the GT for each line image.</p>
POPP Datasets : Datasets for handwriting recognition from French population census
<p><strong>POPP datasets</strong></p> <p>This repository contains 3 datasets created within the POPP project (<a href="https://popp.hypotheses.org/#ancre2">Project for the Oceration of the Paris Population Census</a>) for the task of handwriting text recognition. These datasets have been published in <a href="https://hal.science/hal-03675614/"><em>Recognition and information extraction in historical handwritten tables: toward understanding early 20th century Paris census</em> at DAS 2022.</a></p> <p>The 3 datasets are called “Generic dataset”, “Belleville”, and “Chaussée d’Antin” and contains lines made from the extracted rows of census tables from 1926. Each table in the Paris census contains 30 rows, thus each page in these datasets corresponds to 30 lines.</p> <p>The structure of each dataset is the following:</p> <ul> <li>double-pages : images of the double pages</li> <li>pages: <ul> <li>images: images of the pages</li> <li>xml: METS and ALTO files of each page containing the coordinates of the bounding boxes of each line</li> </ul> </li> <li>lines: contains the labels in the file <code>labels.json</code> and the line images splitted into the folders <em>train</em>, <em>valid</em> and <em>test</em>. The double pages were scanned at a resolution of 200dpi and saved as PNG images with 256 gray levels. The line and page images are shared in the TIFF format, also with 256 gray levels.</li> </ul> <p>Since the lines are extracted from table rows, we defined 4 special characters to describe the structure of the text:</p> <ul> <li>¤ : indicates an empty cell</li> <li>/ : indicates the separation into columns</li> <li>? : indicates that the content of the cell following this symbol is written above the regular baseline</li> <li>! : indicates that the content of the cell following this symbol is written below the regular baseline</li> </ul> <p>We provide a script <code>format_dataset.py</code> to define which special character you want to use in the ground-truth.</p> <p>The split for the <em>Generic Dataset</em> and <em>Belleville</em> have been made at the double-page level so that each writer only appears in one subset among train, evaluation and test. The following table summarizes the splits and the number of writers for each dataset:</p> <table> <thead> <tr> <th>Dataset</th> <th>train - # of lines</th> <th>validation - # of lines</th> <th>test - # of lines</th> <th># of writers</th> </tr> </thead> <tbody> <tr> <td>Generic</td> <td>3840 (128 pages)</td> <td>480 (16 pages)</td> <td>480 (16 pages)</td> <td>80</td> </tr> <tr> <td>Belleville</td> <td>1140 (38 pages)</td> <td>150 (5 pages)</td> <td>180 (6 pages)</td> <td>1</td> </tr> <tr> <td>Chaussée d’Antin</td> <td>625</td> <td>78</td> <td>77</td> <td>10</td> </tr> </tbody> </table> <p> </p> <p><strong>Generic dataset (or POPP dataset)</strong></p> <ul> <li>This dataset is made 4800 annotated lines extracted from 80 double pages of the 1926 Paris census.</li> <li>There is one double page for each of the 80 districts of Paris</li> <li>There is one writer per double page so the dataset contains 80 different writers.</li> </ul> <p> </p> <p><strong>Belleville dataset</strong></p> <p>This dataset is a mono-writer dataset made of 1470 lines (49 pages) from the <em>Belleville</em> district census of 1926.</p> <p> </p> <p><strong>Chaussée d’Antin dataset</strong></p> <p>This dataset is a multi-writer dataset made of 780 lines (26 pages) from the <em>Chaussée d’Antin</em> district census of 1926 and written by 10 different writers.</p> <p> </p> <p><strong>Error reporting</strong></p> <p>It is possible that errors persist in the ground truth, so any suggestions for correction are welcome. To do so, please make a merge request on the <a href="https://github.com/Shulk97/POPP-datasets">Github repository</a> and include the correction in both the labels.json file and in the XML file concerned.</p> <p> </p> <p><strong>Citation Request</strong></p> <p>If you publish material based on this database, we request you to include a reference to paper <a href="http://link.springer.com/chapter/10.1007/978-3-031-06555-2_10"><code>T. Constum, N. Kempf, T. Paquet, P. Tranouez, C. Chatelain, S. Brée, and F. Merveille,Recognition and information extraction in historical handwritten tables: toward understanding early 20th century Paris census ,Document Analysis Systems (DAS), pp. 143- 157, La Rochelle, 2022.</code></a></p>
Unipen data set of on-line (vectorial) handwriting - train_r01_v07
<p>/*****************************************************************************\<br> * *<br> * *<br> * This is the first UNIPEN distribution of the iUF *<br> * *<br> * This distribution comprises NIST train_r01_v07 *<br> * *<br> * http://www.unipen.org/ *<br> * *<br> * Source code: C/Linux at *<br> * http://www.sourcefiles.org/Scientific/Other_Sciences/uptools3.tar.gz *<br> * *<br> * *<br> * The International Unipen Foundation, December 1999 *<br> * *<br> * *<br> *******************************************************************************<br> * *<br> * *<br> * DISCLAIMER AND COPYRIGHT NOTICE FOR ALL DATA CONTAINED ON THIS CDROM: *<br> * *<br> * *<br> * 1) PERMISSION IS HEREBY GRANTED TO USE THE DATA FOR RESEARCH *<br> * PURPOSES. IT IS NOT ALLOWED TO DISTRIBUTE THIS DATA FOR COMMERCIAL *<br> * PURPOSES. *<br> * *<br> * Copyright 1999, International Unipen Foundation - All rights reserved *<br> * *<br> * 2) PROVIDER GIVES NO EXPRESS OR IMPLIED WARRANTY OF ANY KIND AND ANY *<br> * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR PURPOSE ARE *<br> * DISCLAIMED. *<br> * *<br> * 3) PROVIDER SHALL NOT BE LIABLE FOR ANY DIRECT, INDIRECT, SPECIAL, *<br> * INCIDENTAL OR CONSEQUENTIAL DAMAGES ARISING OUT OF ANY USE OF THIS *<br> * DATA. *<br> * *<br> * 4) THE CONDITIONS OF USE REQUIRE PROPER REFERENCE TO THIS DATABASE *<br> * AS DESCRIBED IN ACCOMPANYING DOCUMENT 'unipen-conditions-of-use.html' *<br> * *<br> \*****************************************************************************/</p> <p>Contents of the CDROM:<br> ----------------------</p> <p>1) This file, called CDROM-README<br> 2) The nist distribution, of which part of the directory tree is listed here.</p> <p>train_r01_v07<br> include<br> abm apb app atu bbd ced gmd ibm kai lou pap pri sta uqb<br> aga apc art bba cea cee hpb imp kar mot par rim syn val<br> anj apd ata bbb ceb cef hpp imt lav nic pcl scr tos<br> apa ape att bbc cec dar huj int lex not phi sie ugi</p> <p> data<br> 1a<br> aga apb art ceb gmd imp pri tos val<br> apa app cea ced ibm lou syn uqb<br> 1b 1c 1d 2 3 4 5 6 7 8</p> <p>All files on the the CDROM were tested on UNIPEN integrity using uplib.<br> The description of the contents is given below:</p> <p><br> Description of the contents:<br> ----------------------------</p> <p>For a description and examples of the UNIPEN format, see http://www.unipen.org/</p> <p>The UNIPEN files contained in this release are organized in 10 categories, listed<br> below. The number of .SEGMENTS and number of files for each category are given:</p> <p> cat nsegm nfiles<br> 1a 15953 634 isolated digits<br> 1b 28069 1423 isolated upper case<br> 1c 61351 2145 isolated lower case<br> 1d 17286 1222 isolated symbols (punctuations etc.)<br> 2 122628 2735 isolated characters, mixed case<br> 3 67352 1949 isolated characters in the context of words or texts<br> 4 0 0 isolated printed words, not mixed with digits and symbols<br> 5 0 0 isolated printed words, full character set<br> 6 75529 3298 isolated cursive or mixed-style words (without digits and symbols)<br> 7 85213 3393 isolated words, any style, full character set<br> 8 14544 4563 text: (minimally two words of) free text, full character set</p> <p>In each directory representing a category, e.g., data/1a, a number of<br> sub-directories are contained. The name of a subdirectory is a<br> three-letter word identifying the contributor of the data.</p> <p>Consider for example the UNIPEN files contributed by 'aga' of category<br> 1a (isolated digits). The files containing .SEGMENT entries are contained<br> in the 'data' directory:<br> data/1a/aga</p> <p>Most files in this distribution contain one or more .INCLUDE statements.<br> The corresponding files are found in the 'include' directory, in this case:<br> include/aga<br> Some files (such as the 'imp' contributions) use nested .INCLUDE statements.<br> The software contained in the uptools3 distribution contains code to find<br> files to be included based on an environment variable.</p> <p><br> Distribution of categories per contributor:<br> -------------------------------------------</p> <p> 1a | 1b | 1c | 1d | 2 | 3 | 6 | 7 | 8<br> --------------------------------------------------------------------------------------------------------------<br> abm | | | | | | | 628 4 | 646 4 | 7 3 |<br> aga | 405 14 | 1115 14 | 1063 14 | 221 14 | 2804 14 | | | | 605 14 |<br> anj | | | | | | | 1435 6 | 1435 6 | |<br> apa | 692 74 | 2236 247 | 7414 391 | 1953 268 | 12295 527 | 12295 527 | | | 527 527 |<br> apb | 2033 138 | 3450 466 | 8869 434 | 946 233 | 15298 590 | 15298 590 | | | 590 590 |<br> apc | | | | | | | 1724 441 | 1798 444 | 444 444 |<br> apd | | | | | | | 1958 453 | 2448 507 | 507 507 |<br> ape | | | | | | | 1384 286 | 1848 322 | 322 322 |<br> app | 1046 115 | 3010 353 |10370 556 | 2886 400 | 17312 745 | 17312 745 | | | 745 745 |<br> art | 170 6 | 1042 6 | 2301 6 | 202 6 | 3715 6 | 3715 6 | 687 6 | 933 6 | 186 6 |<br> att | | | | | | | 932 29 | 2253 29 | 819 30 |<br> atu | | | | | | | | | 92 92 |<br> bba | | | | | | | | | 63 63 |<br> bbb | | | | | | | | | 51 51 |<br> bbc | | | | | | | | | 61 61 |<br> bbd | | | | | | | | | 858 858 |<br> cea | 7 3 | 57 6 | 1402 6 | 35 6 | 1501 6 | 1501 6 | 311 6 | 345 6 | 38 6 |<br> ceb | 16 2 | 30 4 | 488 4 | 8 3 | 542 4 | 542 4 | 116 4 | 129 4 | 22 4 |<br> cec | | | | | | | 4880 35 | 5625 35 | 604 35 |<br> ced | 1369 42 | 2691 42 | 2619 43 | 1077 43 | 7756 43 | 7756 43 | | | 1100 43 |<br> cee | | | | | | | 3977 29 | 3978 29 | |<br> dar | | | | | | | 277 2 | 316 2 | 36 2 |<br> gmd | 1145 3 | | 2921 3 | 832 3 | 4898 3 | | | | |<br> hpb | | | | | | | 1524 7 | 2292 7 | 1832 23 |<br> hpp | | | | | | | 8323 32 | 10820 32 | 2591 29 |<br> huj | | | | | | | 104 1 | 104 1 | |<br> ibm | 1571 22 | 4264 22 | 4354 22 | 1994 22 | 12183 22 | | 1196 9 | 1196 9 | |<br> imp | 257 50 | 645 50 | 656 50 | 851 50 | 2409 50 | | 1119 22 | 1119 22 | |<br> imt | | | | | | | 242 1 | 242 1 | |<br> int | | | | | | | 2012 4 | 2012 4 | |<br> kai | | 1961 28 | 8663 46 | 1585 22 | 12209 57 | 8933 28 | 1013 28 | 1663 28 | |<br> kar | | | | | | | 1809 33 | 1860 33 | |<br> lav | | | 1324 9 | | 1324 9 | | 213 5 | 213 5 | |<br> lex | | | | | | | 5660 13 | 7235 13 | 1937 13 |<br> lou | 7 1 | 11 1 | 15 1 | 2 1 | 35 1 | | 1538 7 | 1599 7 | |<br> mot | | | 2701 8 | | 2701 8 | | | | |<br> nic | | | | | | | 6813 66 | 6813 66 | |<br> not | | | | | | | 1452 8 | 1452 8 | |<br> pap | | | | | | | 2203 39 | 2213 41 | |<br> par | | | | | | | 496 8 | 512 8 | |<br> pcl | | | | | | | 616 21 | 616 21 | |<br> phi | | | | | | | 2506 12 | 2506 12 | 91 4 |<br> pri | 78 15 | 212 15 | 191 15 | 230 15 | 711 15 | | 106 3 | 110 3 | 49 18 |<br> rim | | | | | | | 277 21 | 277 21 | |<br> scr | | | | | | | | | 211 44 |<br> sie | | | 377 377 | | 377 377 | | 1593 1593 | 1593 1593 | |<br> sta | | | | | | | 15808 61 | 16415 61 | 156 29 |<br> syn | 4554 17 | 637 8 | 589 8 | 415 8 | 6195 17 | | | | |<br> tos | 543 108 | 1432 108 | 1381 108 | 1660 108 | 4985 108 | | | | |<br> ugi | | | | | | | 597 3 | 597 3 | |<br> uqb | 598 4 | 1514 4 | | 1327 4 | 3439 4 | | | | |<br> val | 1462 20 | 3762 49 | 3653 44 | 1062 16 | 9939 129 | | | | |<br> --------------------------------------------------------------------------------------------------------------<br> | | | | | | | | | |<br> tot |15953 634 |28069 1423|61351 2145|17286 1222|122628 2735|67352 1949 | 75529 3298 | 85213 3393 |14544 4563|<br> --------------------------------------------------------------------------------------------------------------<br> 1a | 1b | 1c | 1d | 2 | 3 | 6 | 7 | 8</p>
ICFHR 2016 Competition on the Classification of Medieval Handwritings in Latin Script - Dataset
<p>The ICFHR2016 Competition on the Classification of Medieval Handwritings in Latin Script (CLaMM), jointly organized by Computer Scientists and Humanists (paleographers) provided a rich database of European medieval manuscripts to the community on Handwriting Analysis and Recognition.</p> <p>If you use this dataset, please cite:</p> <p>Florence Cloppet, Véronique Eglin, Van Cuong Kieu, Dominique Stutzmann, and Nicole Vincent, "ICFHR2016 Competition on the Classification of Medieval Handwritings in Latin Script", <em>Proceedings of International Conference on Frontiers in Handwriting Recognition</em>, Los Alamos : IEEE, 2016, p. 590-595. [<a href="https://doi.org/10.1109/ICFHR.2016.0113">https://doi.org/10.1109/ICFHR.2016.0113</a>]</p> <p>At this competition, we proposed two independent classification tasks which attracted five participants with seven submitted classifiers. Those classifiers are trained on a set of 2000 images with their ground truths. In the first task of script crisp classification, the classifiers have been evaluated on a test set of 1000 single-type manuscripts. In the second task of “Fuzzy Classification”, the classifiers have been carried out on a set of 2000 multi-script-type manuscripts.</p> <p>The present dataset contains the training dataset, both test datasets (task 1 and task 2) and the matrices provided by the competitors. It was first published on a https://clamm.irht.cnrs.fr/icfhr2016-clamm/ in Oct. 2016.</p> <p> </p>
A Serious Game to Anticipate Handwriting Difficulties Screening Through Visual Perception Assessment - DATASET
<p>Each row in the dataset represents a subject. It contains:</p> <ul> <li>The answers to a characterization questionnaire</li> <li>The performance in the game described in the article</li> </ul>
Can Free Drawing Anticipate Handwriting Difficulties? A Longitudinal Study - DATASET
<p>Data to support the conference paper:</p> <p>Dui, L. G., Toffoli, S., Speziale, C., Termine, C., Matteucci, M., & Ferrante, S. (2022, September). Can Free Drawing Anticipate Handwriting Difficulties? A Longitudinal Study. In <em>2022 IEEE-EMBS International Conference on Biomedical and Health Informatics (BHI)</em> (pp. 1-4). IEEE.</p> <ul> <li>BHI22_risk.xlsx: an Excel file with information about: <ul> <li>risk: the risk for handwriting delay, 0=no risk, 1=risk</li> <li>hand: right or left</li> <li>sex: M=male, F=female</li> <li>age: computed at the beginning of the longitudinal study</li> </ul> </li> <li>drawing_features.mat: a Matlab file with five datasets, one for each time point of the longitudinal study, with rows=children, columns=features</li> <li>metadata.mat: features names and type</li> </ul> <p> </p>
Digital Tools for Handwriting Proficiency Evaluation in Children - DATASET
<p>Data to support the findings in the conference paper</p> <p>L. G. Dui, E. Calogero, M. Malavolti, C. Termine, M. Matteucci and S. Ferrante, "Digital Tools for Handwriting Proficiency Evaluation in Children," <em>2021 IEEE EMBS International Conference on Biomedical and Health Informatics (BHI)</em>, Athens, Greece, 2021, pp. 1-4, doi: 10.1109/BHI50953.2021.9508539.</p> <ul> <li>Characterization.xls: an Excel file with sheets: <ul> <li>Characterization: <ul> <li>Subject ID</li> <li>School ID</li> <li>Class</li> <li>Age</li> <li>Sex</li> <li>Hand</li> <li>Years since starting writing in cursive</li> </ul> </li> <li>TabletSUS: score for the System Usability Scale referred to writing on tablet</li> <li>PenSUS: score for the System Usability Scale referred to writing on paper with the smart ink pen</li> <li>BVSCO-2: number of graphemes produced when writing on paper or tablet</li> </ul> </li> <li>pen_tablet_data.mat: a Matlab file with indicators and their names, computed with tablet and pen data</li> </ul>
Data from: High-performance brain-to-text communication via handwriting
Open the record for dataset details and reuse information.
Padeřov-Bible-handwriting-ground-truth: Initial release
<p>This is ground truth based on the Padeřov Bible (Vienna, Austrian National Library, shelfmark Cod. 1175, 1432–1435), the bible of the third redaction of the Old Czech Bible translation. The transcription rules were based on semi-diplomatic transcription rules set by PERO OCR and <em>Směrnice pro vydávání starších českých textů </em>by Jiří Daňhelka (https://vokabular.ujc.cas.cz/moduly/edicnipoznamka.aspx?id=DanhelkaSmernice). Abbreviations were tagged and expanded.</p> <p>Ground truth was created specifically by Anna Michalcová (Czech Academy of Sciences, Czech Language Institute, final check of the transcribed text with the use of her model trained on the Cistercian Bible, New York, The Morgan Library & Museum, shelfmark MS M.752, a.michalcova@ujc.cas.cz), Kamil Bazelides (Comenius University in Bratislava, Faculty of Arts, 4r–8v), Jan Hajič (Czech Academy of Sciences, Masaryk Institute and Archives, 194v–199r), Eliška Pěnkavová (the University of South Bohemia in České Budějovice, Faculty of Arts, 199v–204r), Laura Maniaková (Masaryk University in Brno, Faculty of Arts, 204v–209r), Hana Kreisingerová (Czech Academy of Sciences, Czech Language Institute, 227v–229r), Jitka Filipová (Czech Academy of Sciences, Czech Language Institute, 355r–359v), Chi-hung Lu (Eötvös Loránd University, Faculty of Humanities, 366r–370v) and Martina Dvořáková (Moravian library in Brno, 373r–373v).</p> <p>Produced within the HTR Winter School 2022.</p>
Mathematical Subjective Questions Handwriting Recognition TestSet
Open the record for dataset details and reuse information.
ICDAR 2017 Competition on the Classification of Medieval Handwritings in Latin Script - Dataset
<p>The ICDAR2017 Competition on the Classification of Medieval Handwritings in Latin Script (CLaMM), jointly organized by Computer Scientists and Humanists (paleographers) followed a competition at ICFHR2016 and provided a rich annotated database of European medieval manuscripts to the community on Handwriting Analysis and Recognition, containing information on date of production and class of script.</p> <p>If you use this upload, please cite:</p> <p>Florence Cloppet, Véronique Eglin, Marlène Helias-Baron, van Cuong Kieu, Dominique Stutzmann, Nicole Vincent, "ICDAR 2017 Competition on the Classification of Medieval Handwritings in Latin Script", in <em>14th IAPR International Conference on Document Analysis and Recognition</em>. ICDAR 2017, 1371-76. Kyoto: CPS, 2017. <a href="https://doi.org/10.1109/ICDAR.2017.224">https://doi.org/10.1109/ICDAR.2017.224</a></p> <p>We proposed four independent classification tasks which attracted 10 registered teams, with 6 submitted classifiers from 4 participants. Those classifiers are trained on a set of 3540 images with their ground<br> truths. In task 1 (Script classification) and task 3 (Date classification), the classifiers have been evaluated by a test set of 2000 greyscale, tiff, 300 dpi images. In task 2 (Script classification) and task 4 (Date classification), the test set consists of 1000 images in different formats, resolutions and color<br> representation.</p> <p>The present dataset contains the training dataset, both test datasets (tasks 1 and 3, and tasks 2 and 4) and the matrices provided by the competitors. It was first published on <a href="https://clamm.irht.cnrs.fr/icdar-2017/">https://clamm.irht.cnrs.fr/icdar-2017/</a> in Nov. 2017.</p>
Investigating Visual Perception Impairments through Serious Games and Eye Tracking to Anticipate Handwriting Difficulties - DATASET
<p>In the present dataset, each row represents a subject. For each subject, there are</p> <ul> <li>the ID</li> <li>the gender</li> <li>the class</li> <li>the results in the BVSCO-2 test</li> <li>their position over or under the BVSCO-2 threshold ("prove sopra soglia" represents the number of exercises in which the subject was over the thresold, and "sopra soglia generale" is 1 when a subject is over the threshold in all of the exercises, and 0 otherwise)</li> <li>the features extracted from the game described in the article (times and errors)</li> <li>the features extracted from the data produced by drawing with the Apple Pencil</li> <li>the features extracted from the eye tracker. </li> </ul>
Intensive training programme improves handwriting in a community cohort of people with Parkinson's disease
<p><strong>Background</strong>: People with Parkinson’s disease (PwP) often report problems with their handwriting before they receive a formal diagnosis. Many PwP suffer from deteriorating handwriting throughout their illness, which has detrimental effects on many aspects of their quality of life. </p> <p><strong>Aims:</strong> To assess a 6-week online training programme aimed at improving handwriting of PwP. </p> <p><strong>Methods</strong>: Handwriting samples from a community-based cohort of PwP (n=48) were analysed using Systematic Detection of Writing Problems (SOS-PD) by two independent raters, before and after a 6-week remotely-monitored Physiotherapy-led training programme. Inter-rater variability on multiple measures of handwriting quality was analysed. The handwriting data was analysed using pre/post design in the same individuals. Multiple aspects of the handwriting samples were assessed, including writing fluency, transitions between letters, regularity in letter size, word spacing, and straightness of lines. </p> <p><strong>Results</strong>: Analysis of inter-rater reliability showed high agreement for total handwriting scores, letter size, as well as speed and legibility scores, whereas there were mixed levels of inter-rater reliability for other handwriting measures. Overall handwriting quality (p=0.001) and legibility (p=0.009) significantly improved, while letter size (p=0.012), fluency (p=0.001), regularity of letter size (p=0.009) and straightness of lines (p=0.036) were also enhanced. </p> <p><strong>Conclusions</strong>: The results of this study show that this 6-week intensive remotely-monitored Physiotherapy-led handwriting programme, improved handwriting in PwP. This is the first of its kind study using this tool remotely and it demonstrated that the SOS-PD is reliable for measuring handwriting in PwP.</p> <p><strong>Background</strong>: People with Parkinson’s disease (PwP) often report problems with their handwriting before they receive a formal diagnosis. Many PwP suffer from deteriorating handwriting throughout their illness, which has detrimental effects on many aspects of their quality of life. </p> <p><strong>Aims:</strong> To assess a 6-week online training programme aimed at improving handwriting of PwP. </p> <p><strong>Methods</strong>: Handwriting samples from a community-based cohort of PwP (n=48) were analysed using Systematic Detection of Writing Problems (SOS-PD) by two independent raters, before and after a 6-week remotely-monitored Physiotherapy-led training programme. Inter-rater variability on multiple measures of handwriting quality was analysed. The handwriting data was analysed using pre/post design in the same individuals. Multiple aspects of the handwriting samples were assessed, including writing fluency, transitions between letters, regularity in letter size, word spacing, and straightness of lines. </p> <p><strong>Results</strong>: Analysis of inter-rater reliability showed high agreement for total handwriting scores, letter size, as well as speed and legibility scores, whereas there were mixed levels of inter-rater reliability for other handwriting measures. Overall handwriting quality (p=0.001) and legibility (p=0.009) significantly improved, while letter size (p=0.012), fluency (p=0.001), regularity of letter size (p=0.009) and straightness of lines (p=0.036) were also enhanced. </p> <p><strong>Conclusions</strong>: The results of this study show that this 6-week intensive remotely-monitored Physiotherapy-led handwriting programme, improved handwriting in PwP. This is the first of its kind study using this tool remotely and it demonstrated that the SOS-PD is reliable for measuring handwriting in PwP.</p>
GoBo - A Handwriting Recognition dataset for Personalization
<p>This dataset comprises the images for the personalization described in the paper <em>Personalizing Handwriting Recognition Systems with Limited User-Specific Samples</em>.<br> </p> <p>Dataset Statistics (v.1.0)</p> <p>* Handwritten word-level images<br> * English<br> * 40 Participants<br> * 5 sets from different sources for personalization <br> * 2 sets from 2 domains (same domains as 2 personalization sets) for testing<br> * 926 words/writer, 37k words in total<br> <br> More details can be found on the Github Repository:<br> <a href="https://github.com/catalpa-cl/GoBo/">Github GoBo</a><br> <br> <br> Model<br> gobo_Baselinemodel.hdf5</p>
ScienceDex guides
Understand access before you commit
These curated guides explain access requirements, typical timelines, costs, and reuse considerations for widely used research datasets.
Allen Brain Atlas
Allen Brain Atlas is an Allen Institute collection of brain map atlases, datasets, APIs, and analysis tools covering mouse, human, and non-human primate brain resources.
Annotated Behaviour and Observability Dataset (ABODe)
ABODe is a University of Edinburgh DataShare dataset for behavior classification in group-housed mice using home-cage video, identities, bounding boxes, ground-plate positions, and annotator labels.
DANDI Archive for NWB datasets
DANDI is a BRAIN Initiative archive for publishing and sharing neurophysiology data, including electrophysiology, optophysiology, and behavioral data packaged as NWB and related standards.
International Brain Laboratory public data
The International Brain Laboratory public data releases expose standardized mouse decision-making experiments, including Neuropixels recordings, widefield calcium imaging, behavior, and session metadata accessed through the ONE API.
OpenNeuro
OpenNeuro is a free, open platform for sharing neuroimaging datasets, with public search, dataset pages, and download paths for web, S3, DataLad, and the OpenNeuro CLI.