Talk · International Conference on Electronics, Computer and Computation · June 1, 2023

Kazakh Handwritten Text Recognition Using Computer Vision and Neural Network

Dauletkhan Narynbayev, Ayan Serikkhan, Asmina Barkhandinova, Ilyas Mohammad

Read on dx.doi.orgDOI 10.1109/icecco58239.2023.10147136

Abstract

This paper presents a novel approach to digitizing handwritten text in the Kazakh language. Despite the numerous methods available to recognize handwritten characters, such as Convolution Neural Networks, Recurrent Neural Networks, and Connection Temporal Classification algorithms, these approaches have not been entirely successful in recognizing Kazakh script. Our investigation into the limitations of these existing methods reveals that visual feature extraction alone is insufficient for accurately recognizing Kazakh handwritten text. As a result, we developed a Kazakh Handwritten Text Recognizer (HTR) that combines the strengths of each method. Our experimental results demonstrate that our approach improves the recognition accuracy of Kazakh handwritten text. This paper presents a promising solution for digitizing handwritten Kazakh text and paves the way for further research in this area.

Handwritten Text Recognition TechniquesVehicle License Plate RecognitionHand Gesture Recognition Systemshandwriting recognitionKazakh languagevisual feature extraction

Kazakh Handwritten Text Recognition Using Computer Vision and Neural Network

Abstract

Related publications