Institutional Repository of Vilnius University
Rašysenos atpažinimas ir identifikavimas /
Abstract
dc:descriptionThe project aimed to develop an efficient handwriting recognition model using the Python programming language and various libraries such as os, collections, cv2, pytesseract, numpy, PIL, TensorFlow, and keras. Data was collected from handwritten texts and augmented using data augmentation techniques to increase their quantity and variability. The model architecture included convolutional layers for feature extraction, pooling layers for dimensionality reduction, a reshaping layer to reformat the data, and gated recurrent unit layers for sequence modeling. Model training was conducted using data generators and the ModelCheckpoint callback to achieve optimal model performance. Using the Adam optimizer, the model was efficiently trained and achieved good results. The model learned to accurately recognize Lithuanian letters from images, achieving low training and validation error rates. These results indicate that the model is well-trained and can be effectively used for real-world tasks such as document scanning and text recognition.
Degree
thesis:*- Grantor dc:publisher
- Institutional Repository of Vilnius University
- Year dc:date
- 2024
Author and committee
dc:creator, dc:contributor.*- Author dc:creator
-
- Kaktavičiūtė, Katrina,
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:ELABAETD210648394&prefLang=en_US