Detection of a real Thai national ID card in video based

Name: Aniwat Keereena

Organization : Chulalongkorn University. Department of Computer Engineering

Email : 6470301021@student.chula.ac.th

Name: Sukree Sinthupinyo

Organization : Chulalongkorn University. Department of Computer Engineering

Email : sukree.s@chula.ac.th

LCSH: Identification cards -- Thailand

LCSH: Video surveillance

LCSH: Computer vision

LCSH: Pattern recognition systems

Abstract: Identity verification i s a m andatory p rocess when applying for various online services in Thailand. It requires presentation of your national ID card as evidence. The verification process requires users to capture a photo or video of their card. However, there is a risk of users attempting to counterfeit the card during this process. To mitigate this risk, printed holograms are applied to the card to prevent counterfeiting. The hologram provides a visual means of checking for authenticity. If a hologram is detected on the card, it serves as an indicator of a genuine card, facilitating the decision-making process in identity verification. T his m easure i s i mplemented t o e nhance security and prevent fraudulent activities. This research aims to apply deep learning techniques, utilizing the MobileNet SSD v2 as the backbone model, for hologram recognition on Thai national ID cards. We have developed a system specifically f or classifying holograms using grayscale images. In the training phase, the transfer learning method is employed, transferring pre-trained weights from the RGB model to initialize the grayscale model. The classification a ccuracy o n t he v alidation d ata i ndicates that the system’s mAP (Mean Average Precision) at IoU (Intersection over Union) thresholds of 0.5 to 0.95 for the RGB model outperforms the grayscale model by 1.38%. However, the results for real card detection with grayscale video show an accuracy of 88.33%. These results demonstrate the success of developing a deep learning recognition system for detecting a real Thai national ID card in grayscale with high accuracy.

King Mongkut's University of Technology North Bangkok. Central Library

Address: BANGKOK

Email: library@kmutnb.ac.th

Created: 2024

Modified: 2025-05-26

Issued: 2025-05-26

บทความ/Article

application/pdf

BibliograpyCitation : In IEEE Thailand Section (IEEE Computer Society Thailand Chapter) and Prince of Songkla University. College of Computing. The 21st International Joint Conference on Computer Science and Software Engineering (JCSSE 2024)) (pp.211-216). Phuket : Prince of Songkla University

eng

©copyrights King Mongkut's University of Technology North Bangkok

ลำดับที่.	ชื่อแฟ้มข้อมูล	ขนาดแฟ้มข้อมูล	จำนวนเข้าถึง	วัน-เวลาเข้าถึงล่าสุด
1	JCSSE 2024pp.211-216.pdf	790.94 KB

ใช้เวลา

0.021923 วินาที