Institutional Repository of Vilnius University
Rašysenos atpažinimas ir identifikavimas /
Abstract
dc:descriptionHandwriting is one of human behavioral biometrics and handwriting recognition is a part of computer image analysis. This work covers one area of handwriting recognition writer identification, which is commonly encountered in forensics document expertise. In this work, two Siamese architecture convolutional neural networks are implemented, for the static, text independendent, writer verification task, to answer the question whether the text is written by the same person. The architecture of these networks is inspired by the SigNet and DeepWriter neural networks. Both models were trained on the ”IAM handwriting database”, and two experiments were performed to test and compare their performance. The best results, achieved with one of the networks, were 95.05% verification accuracy with 3.34% equal error rate.
Degree
thesis:*- Grantor dc:publisher
- Institutional Repository of Vilnius University
- Year dc:date
- 2021
Author and committee
dc:creator, dc:contributor.*- Author dc:creator
-
- Karenga, Edvinas,
- Contributors dc:contributor
-
- Grinis, Irus
Rights
dc:rights- Statement dc:rights
-
- info:eu-repo/semantics/openAccess
- Language dc:language
- lit
Identifiers
dc:identifier.*- Repository record dc:identifier
- https://repository.vu.lt/VU:ELABAETD107116560&prefLang=en_US