Deep Convolutional Neural Network Exploiting Transfer Learning for Country Recognition by Classifying Passport Cover

Deep Convolutional Neural Network Exploiting Transfer Learning for Country Recognition by Classifying Passport Cover

Table of Contents




Abstract

Nowadays, Citizen of one country is traveling to another country to settle their various needs through the widespread modern transportation system. However, Passport is a worldly recognized indispensable identity document that is required for traveling internationally. Moreover, Citizen of many countries is strictly prohibited from traveling to other certain countries. So, Passport inspection is a key responsibility for immigration officers in order to confirm the identity of the traveler. In addition to that, it is a laborious and time-consuming task for immigration officers to check all passports meticulously. Hence, automatic country recognition from passport cover images can save a lot of time and physical labor by identifying those unauthorized travelers. Thus in this paper, we have investigated an automatic system using Deep Convolutional Neural Network (DCNN) based on transfer learning with Support Vector Machine (SVM) classifier to analyze passport cover for country identification. Here, the Inception-ResNet-v2 DCNN architecture has been retrained with 80% of image dataset which includes ten classes of passport cover of ten countries using transfer learning method for feature extraction and the extracted feature was then used to train SVM. The proposed model achieved an accuracy level around of 98.75% on the test image dataset.

Author Keywords

  • passport,
  • a deep convolutional neural network,
  • transfer learning,
  • support vector machine,
  • feature extraction

INSPEC: Controlled Indexing

  • convolutional neural nets,
  • feature extraction,
  • image classification,
  • learning (artificial intelligence),
  • public administration,
  • support vector machines

Introduction

In the modern era, global interactions have increased due to the advancement in transportation and communication system. As a consequence, people are now traveling all around the universe to meet up their different purposes such as job, trade, education, treatment, refreshment, and so on [1]. Hence, the number of inward and outward travelers has increased significantly [2]. But, a traveler must have a passport which is an official document certifying the identity and nationality of its holder issued by a country’s government for overseas travel in order to verify their identity and eligibility to enter the country being visited. Besides that, there are many countries whose people are forbidden from entering another country [3]. So, passport inspection is of great importance in order to restrict those illegal travelers. The task of passport inspection is performed by immigration control. The purpose of immigration control is to manage incoming and outgoing travelers through examining their passport. Immigration officer checks the passport by the naked eye that consumes long time in immigration control which is inconvenient for the traveler as well as sometimes it is difficult to find illegal travelers accurately due to inaccurate decision and management [1] [2]. Computer vision and machine learning technologies can be implemented to solve this entire problem. Therefore, we have developed a process for country identification from passport cover utilizing DCNN and SVM to make the process easy and accurate.

Conclusion

In this work, we have proposed an automatic country identification system by assessing the passport cover image using Inception-ResNet-v2 architecture based on transfer learning with SVM classifier.the automatic system helps to identify unauthorized traveler in the immigration control within a short time for overseas travel. Here, the Inception-ResNet-v2 DCNN architecture was used to learn good invariant hidden latent representations from the training image dataset which includes ten classes of passport cover of ten different countries using transfer learning method and the SVM with RBF kernel has performed the recognition task. The proposed model achieved an accuracy level around of 98.75% on the test image dataset. In future fusion and also try to include more classes of passport cover of more countries.

About KSRA

The Kavian Scientific Research Association (KSRA) is a non-profit research organization to provide research / educational services in December 2013. The members of the community had formed a virtual group on the Viber social network. The core of the Kavian Scientific Association was formed with these members as founders. These individuals, led by Professor Siavosh Kaviani, decided to launch a scientific / research association with an emphasis on education.

KSRA research association, as a non-profit research firm, is committed to providing research services in the field of knowledge. The main beneficiaries of this association are public or private knowledge-based companies, students, researchers, researchers, professors, universities, and industrial and semi-industrial centers around the world.

Our main services Based on Education for all Spectrum people in the world. We want to make an integration between researches and educations. We believe education is the main right of Human beings. So our services should be concentrated on inclusive education.

The KSRA team partners with local under-served communities around the world to improve the access to and quality of knowledge based on education, amplify and augment learning programs where they exist, and create new opportunities for e-learning where traditional education systems are lacking or non-existent.

FULL Paper PDF file:

Deep Convolutional Neural Network Exploiting Transfer Learning for Country Recognition by Classifying Passport Cover

Bibliography

author

M. J. Hasan, M. F. Wahid, and M. S. Alom,

Year

2019

Title

Deep Convolutional Neural Network Exploiting Transfer Learning for Country Recognition by Classifying Passport Cover

Publish in

2019 5th International Conference on Advances in Electrical Engineering (ICAEE), Dhaka, Bangladesh, 2019, pp. 870-873,

Doi

10.1109/ICAEE48663.2019.8975666

PDF reference and original file: Click here

 

+ posts

Somayeh Nosrati was born in 1982 in Tehran. She holds a Master's degree in artificial intelligence from Khatam University of Tehran.