Efficient Signature Localization and Detection from Authenticated ID Cards

Read full text Summarise with VIRAT
Authors: B L Thejashwini, H.S. Nagendraswamy
Cited by: 0
References: 1

Summary

No summary yet. VIRAT can read this paper and write a plain-language summary of its aim, methods, key findings and limitations.

Summarise with VIRAT Free to try with your Science Catalogs account.

Abstract

In this digital era, the verification of a person's identity based on authenticated ID cards plays an important role. The verification can be done in many ways; out of that, the signature present on the documents is used for personal authenticity. The main aim of this research article is to localize and detect signatures from various types of authenticated ID cards, such as PAN cards, driving licenses, college IDs, and so on. The deep learning models were used, such as YOLO variants (YOLOv5, YOLOv8, YOLOv9, YOLOv10, YOLOv11), DETR (Detection Transformer), RT-DETR (Real- time Detection Transformer), and Detectron2. The proposed model detects the signature from newly collected datasets, some images sourced from the MASK-RCNN-Dataset and the Roboflow platform. The model was tested with our newly collected datasets without any preprocessing. To measure the effectiveness of the models, we have used the most common metric in detection phase as, mean average precision (mAP50 and mAP50-95). All the models detect signatures with above 96% accuracy. This article also compares all the models and observes that the RT-DETR model gives efficient results. Specifically, using the RT- DETR_r101vd_coco_0365 backbone model, RT-DETR achieved a detection accuracy of 100% with a pre-processed dataset and also observed that all models obtained an IoU ranging from 92% to 97%.

Fields of Study & Subjects

Handwritten Text Recognition Techniques Advanced Neural Network Applications Vehicle License Plate Recognition

Publication details

Journal / source
International Journal of Advances in Signal and Image Sciences
Volume
12
Issue
3s
Pages
726–736
Publication date
1 January 2026
Language
EN