Hyperspectral imaging for Ink Identification in Handwritten Documents

سلطانی, پیمان and پناهی, مجید and درودی, احمد (2019) Hyperspectral imaging for Ink Identification in Handwritten Documents. In: 1st International Conference on Optoelectronics, Applied Optics and Microelectronics - 2019.

Text 1 (53).pdf Download (1MB)

Abstract

In order to correctly enforce the law in valid documents, intrusion detection has become an important factor in identifying fake documents. In this paper, an approach using Hyperspectral Imaging and Support-vector machine classification (SVM) is used to ink Analysis in handwritten documents and pen type test. The method used gives the results through a careful study of the properties of the ink and the pen without touching the sample. We use a classifier based on the inherent properties of the ink that derives from its physical properties. In addition, a database of hyperspectral curves is made up of several ink types that are used to obtain the properties of different inks. The proposed method for Ink Identification using seven different pens and more than 200 samples has been tested and the results are presented.

Actions (login required)

View Item