Engraved digit detection using HOG-real AdaBoost and deep neural network

Authors

TUAN LINH DANG
THANG CAO
YUKINOBU HOSHINO

DOI

10.3906/elk-2001-12

Abstract

This paper proposes a framework for recognizing sequences of digits engraved on steel plates. These digits are normally blurred, dirty, not clear, tilted, and sometimes overlapped by other digits. Several digits in a string with uneven spacing and different sizes are detected at the same time. The framework consists of two main components called histogram of oriented gradient-real AdaBoost module and deep neural network module. The first component is used to detect digit windows, and the second component is employed to recognize digits inside the detected windows. Experimental results demonstrated that the proposed framework could be a potential solution to recognize the engraved digits.

Keywords

Histogram oriented gradient, real AdaBoost, deep neural network, engraved digit recognition

First Page

138

Last Page

151

Recommended Citation

DANG, TUAN LINH; CAO, THANG; and HOSHINO, YUKINOBU (2021) "Engraved digit detection using HOG-real AdaBoost and deep neural network," Turkish Journal of Electrical Engineering and Computer Sciences: Vol. 29: No. 1, Article 10. https://doi.org/10.3906/elk-2001-12
Available at: https://journals.tubitak.gov.tr/elektrik/vol29/iss1/10