Comparison for Handwritten Character Recognition and Handwritten Text Recognition and Tesseract Tool on IJAZAh�s Handwriting

Setiawan, Alexander and Gunadi, Kartika and , MADE YOGA MAHARDIKA (2023) Comparison for Handwritten Character Recognition and Handwritten Text Recognition and Tesseract Tool on IJAZAh�s Handwriting. [UNSPECIFIED]

	PDF Download (1135Kb)
	PDF Download (1483Kb)

Abstract

Handwriting is a form of being able to recognize various types of writing in various existing fonts. Unlike consistent computer letters, each human handwriting is unique in its form and consistency. These problems can be found in a document where the data is in the form of handwriting. Segmentation of the data location will use a run length smoothing algorithm with points as segmentation features. The Handwriting Text Recognition (HTR) technique requires segmented data intowords. The HandwritingCharacter Recognition (HCR) technique requires segmented data into various characters. The process of this HCR technique uses the LeNet5 model using the EMNIST dataset. HTR uses the tesseract tool and a convolutional iterative neural network using the IAM database. Experiment on 10 samples of scan images, segmentation obtained an average accuracy of 95.6%. The HCR technique failed in the letter segmentation process in cursive handwriting. The easiest technique to use is the HTR with the helps of tesseract tool, tesseract tool also has a good performance. Tesseract managed to get word accuracy above 70% tested on 5 scan samples, 15 data fields.

Actions (login required)