• Users Online: 205
  • Home
  • Print this page
  • Email this page
Home About us Editorial board Ahead of print Current issue Search Archives Submit article Instructions Subscribe Contacts Login 


 
 Table of Contents  
ORIGINAL ARTICLE
Year : 2017  |  Volume : 3  |  Issue : 4  |  Page : 210-216

Hyperspectral imaging technology for revealing the original handwritings covered by the same inks


1 Key Laboratory of Evidence Science, China University of Political Science and Law, Ministry of Education; 2011 Collaborative Innovation Center of Judicial Civilization, Beijing 100088; Fada Institute of Forensic Medicine and Science, Beijing 100192, China
2 The Ministry of Public Security Material Evidence Identification Center, Beijing 100038, China

Date of Web Publication11-Jan-2018

Correspondence Address:
Dr. Yuanyuan Lian
Institute of Evidence Law and Forensic Science, China University of Political Science and Law, Beijing
China
Login to access the Email id

Source of Support: None, Conflict of Interest: None


DOI: 10.4103/jfsm.jfsm_77_17

Rights and Permissions
  Abstract 

This manuscript presents a preliminary investigation on the applicability of hyperspectral imaging technology for nondestructive and rapid analysis to reveal covered original handwritings. The hyperspectral imager Nuance-Macro was used to collect the reflected light signature of inks from the overlapping parts. The software Nuance1p46 was used to analyze the reflected light signature of inks which shows the covered original handwritings. Different types of black/blue ballpoint pen inks and black/blue gel pen inks were chosen for sample preparation. From the hyperspectral images examined, the covered original handwritings of application were revealed in 90.5%, 69.1%, 49.5%, and 78.6% of the cases. Further, the correlation between the revealing effect and spectral characteristics of the reflected light of inks at the overlapping parts was interpreted through theoretical analysis and experimental verification. The results indicated that when the spectral characteristics of the reflected light of inks at the overlapping parts were the same or very similar to that of the ink that was used to cover the original handwriting, the original handwriting could not be shown. On the contrary, when the spectral characteristics of the reflected light of inks at the overlapping parts were different to that of the ink that was used to cover the original handwriting, the original handwriting was revealed.

Keywords: Covered original handwritings, hyperspectral imaging technology, metamerism, revealing


How to cite this article:
Lian Y, Liang L, Li B. Hyperspectral imaging technology for revealing the original handwritings covered by the same inks. J Forensic Sci Med 2017;3:210-6

How to cite this URL:
Lian Y, Liang L, Li B. Hyperspectral imaging technology for revealing the original handwritings covered by the same inks. J Forensic Sci Med [serial online] 2017 [cited 2018 Jul 18];3:210-6. Available from: http://www.jfsmonline.com/text.asp?2017/3/4/210/222895


  Introduction Top


In practice, documents cannot be used to prove facts in several cases because the date, amount, and other important contents have been blotted out either intentionally or unintentionally. Therefore, in the examination of questioned documents, revealing the original handwritings is particularly important.

In those cases, the original handwriting is not easy to observe because it has been covered by opaque materials such as writing inks.[1] There are two types of covered original handwritings: those covered by heterogeneous materials, such as blue gel pen ink handwriting covered with blood, paint, or other pigment and those covered by the same type of materials, such as blue gel pen ink handwriting covered with black gel pen inks.[1] The original handwriting covered by the same type of materials can be divided into that covered by inks of the same type and color (blue ballpoint pen handwriting covered with another blue ballpoint ink) and that covered by inks of the same type but not the same color (blue ballpoint pen handwriting covered with black ballpoint ink). In the field of questioned document examination, it has been difficult to reveal the original handwriting covered by inks of the same type and color because the color, composition, and related physical and chemical properties of the inks of the original handwriting and covering layer are very similar to each other. According to the literature, there are several methods for revealing the covered original handwriting, such as removal of surface coating, chemical processes, and electrostatic imaging tests.[2],[3] These methods have shown strong analytical performance for solving many cases regarding covered handwritings. Some of them are more informative and objective than others. Each of them has their own advantages and limitations. However, they fail in a great number of cases, especially when two very similar ink lines are present. In this case, a nondestructive, high-sensitivity, reliable method to reveal the covered original handwritings is one of the goals of questioned document examiners.

In the recent years, the spectral imaging technique has been applied to studies in the field of forensic science,[4],[5],[6],[7],[8],[9] such as those concerning fingerprints, paint, and fiber. With respect to handwriting, the spectral characteristics vary owing to the diversity in ink composition, namely, the reflection (absorption) curves show differences, although they look the same on the surface in color.[6] The metamerism of inks with the same color offers the possibility of revealing the original handwritings by using the spectral imaging technique.

In this work, the principle and effect of using the hyperspectral imaging technology to reveal original handwriting covered by inks of the same type and color is studied through a series of experiments, in order to explore a new nondestructive method with a high detection rate. At the same time, we hope that these findings could promote the application of hyperspectral imaging technology in the field of forensic science.


  Experiments Top


Sampling

A total of 22 black ballpoint pens, 62 blue ballpoint pens, 42 black gel pens, and 62 blue gel pens of different brands were purchased in the domestic market of China. The samples were prepared in the following manner by the same person: pens of the same type and color were used as the original handwriting and covering painting pairwise as shown in [Figure 1]. All combinations were prepared with these pens, and each combination was additionally prepared in the opposite permutation, giving rise to a total of 462, 2782, 2782, and 1722 combinations to the corresponding pens, respectively. The same office copy/print white paper (70 g/m2) was always used. In this study, 24 h was the time lapse between the applications of handwriting and covering painting inks. All inks were allowed to dry at least for a few hours under the same ambient laboratory conditions before the analyses.
Figure 1: Red, green, and blue image of the sample using black ballpoint pen 16 covering black ballpoint pen 1

Click here to view


Instrumentation

The samples were examined with a Cambridge Research and Instrumentation Nuance-Macro hyperspectral imager with Nuance1p46 software (manufactured by Cambridge Research and Instrumentation, USA). A metal halide lamp was the light source. The wavelength range of the image acquisition set was from 420 to 720 nm, with a step size between successive acquisitions of 10 nm. The exposure model was automatic. The method used to analyze the image and data set consisted of removing the mixture.


  Results and Discussions Top


As shown in [Figure 2], theoretically, there are four layers in the covered handwritings: the layer of paper (named A), the layer of the original handwriting (named B), the layer of the ink used to cover the original handwriting (named C), and the layer of the overlapping part mixing B with C (named D).
Figure 2: Sketch of the covered original handwritings

Click here to view


The colors of layers C and D are similar under the naked eye, and thus, it is difficult to identify the original handwriting. From the optical point of view [Figure 3], the reflected light c is the reflection of incident light by layer C; the reflected light b is the reflection of the transmission light from layer C by layer B; and the reflected light a is the reflection of the transmission light from layer B by layer A. That is to say, the reflected light d, which is the signature collected by the hyperspectral imager at the layer of the overlapping part D, is the mixture of lights a, b, and c. Several factors could determine the formation of the reflected light d as follows: the amount of ink applied, which mostly depends on writing pressure and the type of point and pen used; the chemical composition of the ink, which is patented and usually unknown to the forensic examiner; the time that separates the application of both inks, which is directly related to the drying processes of the inks; the structure and surface of the paper substrate, which determine the hydraulic conductivity and absorption of the ink; and climatic factors such as heat, humidity, and light, which may also play a role. In this study, the chemical composition of the ink, which determines the absorption, transmission, and reflectance properties of each layer, was studied. Other factors, such as the type of paper and drying times, should be studied further in the future.
Figure 3: Sketch of the formation of the reflected light at the layer of overlapping part

Click here to view


Revealing effects

It is difficult to identify the covered handwritings under the naked eye. During the process of hyperspectral imaging analysis, feature points were taken on layers A, B, C, and D. The four layers in the images were given a color. White, red, yellow, and magenta colors were given to the layers A, B, C, and D, respectively. The hyperspectral image shows the original handwritings clearly, as shown in [Figure 4],[Figure 5],[Figure 6],[Figure 7].
Figure 4: (a) Red, green, and blue image of the sample using black ballpoint pen 2 covering black ballpoint pen 1. (b) Hyperspectral image of the sample using black ballpoint pen 2 covering black ballpoint pen 1. (c) Red, green, and blue image of the sample using black ballpoint pen 11 covering black ballpoint pen 10. (d) Hyperspectral image of the sample using black ballpoint pen 11 covering black ballpoint pen 10

Click here to view
Figure 5: (a) Red, green, and blue image of the sample using blue ballpoint pen 16 covering blue ballpoint pen 18. (b) Hyperspectral image of the sample using blue ballpoint pen 16 covering blue ballpoint pen 18. (c) Red, green, and blue image of the sample using blue ballpoint pen 57 covering blue ballpoint pen 40. (d) Hyperspectral image of the sample using blue ballpoint pen 57 covering blue ballpoint pen 40

Click here to view
Figure 6: (a) Red, green, and blue image of the sample using black gel pen 25 covering black gel pen 27. (b) Hyperspectral image of the sample using black gel pen 25 covering black gel pen 27. (c) Red, green, and blue image of the sample using black gel pen 31 covering black gel pen 35. (d) Hyperspectral image of the sample using black gel pen 31 covering black gel pen 35

Click here to view
Figure 7: (a) Red, green, and blue image of the sample using blue gel pen 9 covering blue gel pen 11. (b) Hyperspectral image of the sample using blue gel pen 9 covering blue gel pen 11. (c) Red, green, and blue image of the sample using blue gel pen 17 covering blue gel pen 39. (d) Hyperspectral image of the sample using blue gel pen 17 covering blue gel pen 39

Click here to view


As an exploratory study, the hyperspectral images were divided into two classes: visible and latent, according to whether the covered original handwritings could be identified or not by people who had never seen the original handwriting with the naked eye, as shown in [Figure 8].
Figure 8: (a) Visible hyperspectral images. (b) Invisible hyperspectral images

Click here to view


The statistics of all the hyperspectral images are shown in [Table 1], according to the aforementioned classification standard.
Table 1: Statistics of the hyperspectral imaging results on the covered original handwritings

Click here to view


These hyperspectral images and statistics showed that when the covered original handwritings were examined by the hyperspectral imaging technology, it could give us a direct observation of the covered original handwritings and a higher visible rate. The order of visible rate from high to low was as follows: black ballpoint pen, blue ballpoint pen, blue gel pen, and black gel pen.

Correlations between spectroscopic properties of overlapping parts and hyperspectral images

The spectroscopic properties of the overlapping parts have been studied to understand the hyperspectral images and their influence in revealing the covered original handwritings.

In this part of the study, four typical combinations of pens of the same type and color were chosen to prepare the testing samples in the following manner, taking black gel pens 31 and 35 as an example. The first ink (C31) was always a vertical line applied from top to bottom, the second ink (C35) line crossed horizontally the first from left to right, and the first ink (C31) crossed again vertically the second from top to bottom [Figure 9]a. The feature points were taken on the first ink (C31) and second ink (C35). Red and yellow colors were given to the first ink (C31) and second ink (C35), respectively. The hyperspectral image reveals the different results of the first (C31) and second (C35) inks, as shown in [Figure 9]b.
Figure 9: (a) Red, green, and blue image of the sample using black gel pen 31 and black gel pen 35. (b) Hyperspectral image of the sample using black gel pen 31 and black gel pen 35

Click here to view


Then, feature points were taken on the first ink (C31), second ink (C35), left crossing (the second ink [C35] covered the first ink [C31]), and right crossing (the first ink [C31] covered the second ink [C35]). The four points in the images were given a color. The red, yellow, green, and blue colors were given to the first ink (C31), second ink (C35), left crossing, and right crossing, respectively. The hyperspectral image shows the different results of the four points, as shown in [Figure 10].
Figure 10: Hyperspectral images of the crossings using black gel pen 31 and black gel pen 35

Click here to view


The results show that neither the left crossing nor the right crossing share the same hue as the inks applied. Meanwhile, the hue of the left crossing was different from that of the right crossing, owing to the opposite sequence. The spectroscopic properties of the crossings were related to the revealing results of the covered handwritings between the two inks. Both the hyperspectral images of covered handwritings using black gel pen 31 and black gel pen 35 [Figure 11] offered us the direct observation of the covered original handwritings. Meanwhile, the hue of the overlapping layer was different from either the hue of the original handwritings or the hue of the inks that covered the original handwritings, independent of the sequence.
Figure 11: Hyperspectral images of the covered samples using black gel pen 31 and black gel pen 35

Click here to view


Regarding the black gel pen 25 and black gel pen 27, there were significant differences between the two inks applied [Figure 12].
Figure 12: (a) Red, green, and blue image of the sample using black gel pen 25 and black gel pen 27. (b) Hyperspectral image of the sample using black gel pen 25 and black gel pen 27

Click here to view


The hyperspectral images of the crossings show that the left crossing has a different hue from that of the black gel pen 25, but it shares the same hue as the black gel pen 27. The hue of the right crossing differed from either of the applied inks [Figure 13]. Owing to the spectroscopic properties of the crossings, the covered handwriting of black gel pen 25 over black gel pen 27 was visible, but it was invisible with the opposite sequence [Figure 14].
Figure 13: Hyperspectral images of the crossings using black gel pen 25 and black gel pen 27

Click here to view
Figure 14: Hyperspectral images of the covered samples using black gel pen 25 and black gel pen 27

Click here to view


Regarding the black gel pen 47 and black gel pen 50, there were some differences between the two inks applied [Figure 15].
Figure 15: (a) Red, green, and blue image of the sample using black gel pen 47 and black gel pen 50. (b) Hyperspectral image of the sample using black gel pen 47 and black gel pen 50

Click here to view


The hyperspectral images of both crossings shared the same hue as the black gel pen 50, but differed from the black gel pen 47 [Figure 16]. The covered handwriting of black gel pen 47 over black gel pen 50 was visible. However, it was invisible with the opposite sequence [Figure 17] because of the spectroscopic properties of the crossings.
Figure 16: Hyperspectral images of the crossings using black gel pen 47 and black gel pen 50

Click here to view
Figure 17: Hyperspectral images of the covered samples using black gel pen 47 and black gel pen 50

Click here to view


Regarding the blue gel pen 8 and blue gel pen 12, there were no differences between the two inks applied [Figure 18].
Figure 18: (a) Red, green, and blue image of the sample using blue gel pen 8 and blue gel pen 12. (b) Hyperspectral image of the sample using blue gel pen 8 and blue gel pen 12

Click here to view


The hyperspectral images showed that both crossings and the two inks applied shared the same blend color of the four colors [Figure 19]. For this reason, neither of the covered handwritings was visible [Figure 20].
Figure 19: Hyperspectral images of the crossings using blue gel pen 8 and blue gel pen 12

Click here to view
Figure 20: Hyperspectral images of the covered samples using blue gel pen 8 and blue gel pen 12

Click here to view


The aforementioned results showed that the visibility of the hyperspectral images of the covered handwritings corresponds with the spectroscopic properties of the overlapping. For revealing the covered handwritings by the hyperspectral imaging technology, it was necessary that the overlapping part and the inks covering the original handwritings must show different spectra and the differences should be detectable to the hyperspectral imager.

Summary

Based on the theoretical analysis and experimental verification in Section 3.2, the differences among the reflection, absorption, and transmission properties of the inks, especially the inks covering the original handwritings, were the main reason why the visible rates of different types of pens varied.

In hyperspectral images, the colorants (dyes and pigments) mainly affect the optical properties of the inks; thus, the signatures detected vary with the different types of colorants. In this case, there may be several pens, whose inks show similar spectra and no differences were identified; therefore, the covering combinations involving these pens will reduce the visible rate.

For ballpoint pens, the ink colorants are usually a mixture of a variety of dyes that have high transmission properties, and thus, the visible rate of the covered handwritings with black ballpoint pens is as high as 90.5%. Compared with the black ballpoint pens, the colorants of the blue ballpoint pen inks are also a mixture of dyes, but there are not as many types of blue dyes to apply in the inks. Thus, there is a higher proportion of invisible covering combinations. The visible rate of coverings using blue ballpoint pens is significantly lower than the rate of coverings using black ballpoint pens, <70%.

Concerning writing pens, when an ink line is applied, it disperses throughout the paper and is absorbed into it. When the second ink is applied over another, it can disperse across the void spaces as well as the other ink. Oil-based inks (ballpoint pens) tend toward being only partially absorbed, and most of the ink stays adhered to the surface of the paper, owing to its high viscosity. Water-based inks (gel pens) tend to soak into the paper fibers, like water into a sponge. For this reason, although the blue gel pens share similar dyes to those of the blue ballpoint pens, the transmission properties of the blue gel ink layers are higher, owing to the thinner layers. Thus, the visible rate of the coverings using blue gel pens is close to 80%, higher than the blue ballpoint pens.

Regarding the black gel pens, the ink colorants are dyes, pigments, and toner. The handwritings are invisible when covered by the toners because they exhibit very strong absorption of light. In this study, there are 21 black gel pens, among the 62 black gel pens, whose inks contain toner. Therefore, the visible rate of the coverings using black gel pens is the lowest, only about 40%.


  Conclusions Top


This manuscript has presented the potential of hyperspectral imaging technology to reveal the covered handwritings with the application of pens of the same ink type and color. The advantage of hyperspectral imaging is that it provides discriminating optical information on the overlapping part of the inks in a nondestructive and rapid way. The colorful figures allowed direct visualization of the original handwritings that facilitated the identification of the covered handwritings with the same type and color of ink. It was also found that its ability to reveal the coverings depends on the spectroscopic properties of the overlapping part. When there are different spectra, and thus, differences between the overlapping part and the inks covering the original handwritings, it is possible to reveal the covered handwritings by hyperspectral imaging technology.

The influence of the other factors, such as the amount of ink applied, time that separates the application of both inks, and structure and surface of the paper substrate, will be studied in the future.

Financial support and sponsorship

This study was supported by Program for Young Innovative Research Team in China University of Political Science and Law (1000-10814344) and Young Research in China University of Political Science and Law (16ZFQ82008).

Conflicts of interest

There are no conflicts of interest.

 
  References Top

1.
Jiantong H. Questioned Document Examination. Beijing: Publishing House of CPPSU; 2003. p. 151-5, 172-198.  Back to cited text no. 1
    
2.
Guiqiang W. Optical and Photographic Techniques on Fingerprints. Beijing: Masses Publishing House; 2001. p. 125-411.  Back to cited text no. 2
    
3.
Guiqiang W. The introduction of hyperspectral imaging technology. J Forensic Sci Technol 2004;1:7-12.  Back to cited text no. 3
    
4.
Wei H, Guiqiang W. Research on the application of hyperspectral imaging technology in forensic science. J Forensic Sci Technol 2006;2:1-4.  Back to cited text no. 4
    
5.
Wei H. Research on the Application of Hyperspectral Imaging Technology in Forensic Science. Master's Thesis. Beijing: Department of the CPPSU; 2003.  Back to cited text no. 5
    
6.
Peng N. Research on the Identification of ink Types Using Hyperspectral Imaging Technology. Master's Thesis. Beijing: Department of the CPPSU; 2008.  Back to cited text no. 6
    
7.
Exline DL, Schuler R, Treado PJ. Improved fingerprint visualization using luminescence and visible reflectance chemical imaging. Forensic Sci Commun 2003;5;35-46.  Back to cited text no. 7
    
8.
Crane NJ, Bartick EG, Perlman RS, Huffman S. Infrared spectroscopic imaging for noninvasive detection of latent fingerprints. J Forensic Sci 2007;52:48-53.  Back to cited text no. 8
    
9.
Braz A, López-López M, García-Ruiz C. Raman imaging for determining the sequence of blue pen ink crossings. Forensic Sci Int 2015;249:92-100.  Back to cited text no. 9
    


    Figures

  [Figure 1], [Figure 2], [Figure 3], [Figure 4], [Figure 5], [Figure 6], [Figure 7], [Figure 8], [Figure 9], [Figure 10], [Figure 11], [Figure 12], [Figure 13], [Figure 14], [Figure 15], [Figure 16], [Figure 17], [Figure 18], [Figure 19], [Figure 20]
 
 
    Tables

  [Table 1]



 

Top
 
 
  Search
 
Similar in PUBMED
   Search Pubmed for
   Search in Google Scholar for
 Related articles
Access Statistics
Email Alert *
Add to My List *
* Registration required (free)

 
  In this article
Abstract
Introduction
Experiments
Results and Disc...
Conclusions
References
Article Figures
Article Tables

 Article Access Statistics
    Viewed506    
    Printed16    
    Emailed0    
    PDF Downloaded89    
    Comments [Add]    

Recommend this journal


[TAG2]
[TAG3]
[TAG4]