Handwritten Digit Recognition Using Convolutional Neural Networks and Big Data Processing

Keywords:
Machine Learning, Big Data, Digit Recognition, Artificial Intelligence

Abstract:
In this paper, we present a handwriting recognition system that combines neural networks and big data processing. We use a custom convolutional model, and optimized data augmentation techniques. To manage and process large volumes of data more efficiently, we relied on Apache Spark. Additionally, a user-friendly API has been created to enable real-time recognition of handwritten digits. Evaluation of the system on a custom dataset shows extremely high accuracy, with precision greater than 98% on the test data. The main challenge now is to process large datasets, as well as manage different handwriting styles and ensure that the models perform accurately in realworld scenarios. The main contributions of this work are implementations of an efficient convolutional model for recognizing handwritten digits, a system for processing big data in a distributed environment, web API and user interface for real-time handwriting recognition, model error analysis and suggestions for further improvements.

CITATION:

IEEE format

Copy to clipboard

P. Matijašević, M. Mravik, “Handwritten Digit Recognition Using Convolutional Neural Networks and Big Data Processing,” in Sinteza 2025 - International Scientific Conference on Information Technology, Computer Science, and Data Science, Belgrade, Singidunum University, Serbia, 2025, pp. 531-535. doi:10.15308/Sinteza-2025-531-535

APA format

Copy to clipboard

Matijašević, P., Mravik, M. (2025). Handwritten Digit Recognition Using Convolutional Neural Networks and Big Data Processing. Paper presented at Sinteza 2025 - International Scientific Conference on Information Technology, Computer Science, and Data Science. doi:10.15308/Sinteza-2025-531-535

BibTeX format
Download

RefWorks Tagged format
Download