A CNN Framework for Real-Time and Edge Deployment with High Accuracy for Lightweight Deep Learning for Bengali OCR
International Journal on Smart & Sustainable Intelligent Computing · 18.04.2026 · 3 (1) · 25-50
DOI: 10.63503/j.ijssic.2026.239 ↗Kurz gefasst
Fachartikel in der Zeitschrift International Journal on Smart & Sustainable Intelligent Computing, online erschienen am 18.04.2026. Verfasst von Dipankar Dey und 1 weiteren Autor:innen.
Abstract (Originalsprache)
The Bengali script, like other Indic scripts, has a large character set, complex conjunct characters, and varying handwriting styles, making optical character recognition (OCR) a complex task. Although convolutional neural networks (CNNs), ensemble techniques, and CNN-Transformer models have shown high accuracy, their high computational complexity and large input size make them unsuitable for mobile and embedded systems. To make the processing cost-effective and improve the accuracy of handwritten Bengali character recognition, this paper proposes a light-weight CNN with optimized filters and a 40 × 40 input size. Compared to the existing deep learning and hybrid approaches, the proposed approach has shown 98.29\% top-1 accuracy on a 50-class dataset with 12,000 training samples and 2,997 testing samples, reducing the parameter size and computational complexity. A brief literature review is also presented to discuss the evolution of real-time and large-scale Bengali OCR systems and the challenges involved in compound character recognition, efficiency, and system-level validation.
Zitieren
APA
Dey, D., & Jana, D. K. (2026). A CNN Framework for Real-Time and Edge Deployment with High Accuracy for Lightweight Deep Learning for Bengali OCR. International Journal on Smart & Sustainable Intelligent Computing, 3(1), 25-50. https://doi.org/10.63503/j.ijssic.2026.239
Vancouver
Dey D, Jana DK. A CNN Framework for Real-Time and Edge Deployment with High Accuracy for Lightweight Deep Learning for Bengali OCR. International Journal on Smart & Sustainable Intelligent Computing. 2026;3(1):25-50. doi:10.63503/j.ijssic.2026.239
BibTeX
@article{Dey2026,
author = {Dey, Dipankar and Jana, Dipak Kumar},
title = {A CNN Framework for Real-Time and Edge Deployment with High Accuracy for Lightweight Deep Learning for Bengali OCR},
journal = {International Journal on Smart & Sustainable Intelligent Computing},
year = {2026},
volume = {3},
number = {1},
pages = {25--50},
doi = {10.63503/j.ijssic.2026.239}
}
EUVIMED