Results 101 to 110 of about 42,098,161 (146)

Nuremberg Letterbooks: A Multi-Transcriptional Dataset of Early 15th Century Manuscripts for Document Analysis. [PDF]

Sci Data
Mayr M, Krenz J, Neumeier K, Bub A, Bürcky S, Brolich N, Herbers K, Habermann M, Fleischmann P, Maier A, Christlein V.   +10 more
europepmc   +1 more source

Leveraging OCR and HTR cloud services towards data mobilisation of historical plant names. [PDF]

Int J Digit Humanit
Sadek J, Vlachidis A, Pickering V, Humbel M, Metilli D, Carine M, Nyhan J.   +6 more
europepmc   +1 more source

Integrating CNN and transformer architectures for superior Arabic printed and handwriting characters classification. [PDF]

Sci Rep
Al-Maamari MR, Ramteke R, Al-Hejri AM, Alshamrani SS.   +3 more
europepmc   +1 more source

Deep Learning for Historical Document Analysis and Recognition-A Survey. [PDF]

J Imaging, 2020
Lombardi F, Marinai S.
europepmc   +1 more source

iForal: Automated Handwritten Text Transcription for Historical Medieval Manuscripts. [PDF]

J Imaging
Matos A, Almeida P, Correia PL, Pacheco O.   +3 more
europepmc   +1 more source

OICEN-HTR: Old Icelandic / Norse Handwritten Text Recognition Test Upload

This repository contains OICEN Core v.0.3, an HTR model fine-tuned with CATMuS Medieval 1.6.0 on Old Icelandic manuscripts. In-domain test character accuracy is 0.94 and word accuracy is 0.82. Out-of-domain test character / word accuracy: - Kringla-Fragment: 0.86 / 0.56 - Part of Codex Frisianus: 0.81 / 0.46 - DG 4-7 (Pamphilus saga): 0.88 / 0.66 - NRA
openaire   +1 more source

A Pix2Pix Architecture for Complete Offline Handwritten Text Normalization. [PDF]

Sensors (Basel)
Barreiro-Garrido A, Ruiz-Parrado V, Moreno AB, Velez JF.   +3 more
europepmc   +1 more source

A Computer Vision Framework for Structural Analysis of Hand-Drawn Engineering Sketches. [PDF]

Sensors (Basel)
Joffe I, Qian Y, Talebi-Kalaleh M, Mei Q.   +3 more
europepmc   +1 more source

OICEN-HTR: Old Icelandic / Norse Handwritten Text Recognition (PP-OCRv6-medium)

This deposition contains OICEN-HTR (PP-OCRv6-medium version), an HTR model fine-tuned with Kraken's PP-OCRv6-medium model on Old Icelandic / Norse manuscripts and fragments. In-domain held-out test character accuracy is 90.77% and word accuracy is 72.98%. Out-of-domain test character / word accuracy: * Kringla-Fragment Lbs.
openaire   +1 more source

The digitisation workflow of the herbarium of the State Museum of Natural History of the NAS of Ukraine (LWS). [PDF]

Biodivers Data J
Novikov A, Nachychko V.
europepmc   +1 more source