First investigation of feasibility of contact-less non-destructive optical sensors to detect, acquire and digitally process forensic handwriting based on pressure information | IEEE Conference Publication | IEEE Xplore

First investigation of feasibility of contact-less non-destructive optical sensors to detect, acquire and digitally process forensic handwriting based on pressure information

Publisher: IEEE

Abstract:

Both in biometric and forensic sciences, handwriting modality plays an important role. While in biometrics, the sub-discipline of online analysis utilises recordings of t...View more

Abstract:

Both in biometric and forensic sciences, handwriting modality plays an important role. While in biometrics, the sub-discipline of online analysis utilises recordings of the actual writing process, offline analysis has a close link to handwriting forensics. In the later domain, goals comprise for example writer identification or uncover indented writing of found writing traces on documents. In this paper we study the general feasibility of a new approach to handwriting forensics by means of three contact-less non-destructive 2D/3D optical sensors providing high resolution images in the nanometer range which have already shown great potential for the digitised crime scene forensics e.g for fingerprints or gun cartridges. We perform a three-step concept: First, in a general sensor review we identify one sensor which appears most appropriate for our initial use cases, along with appropriate parametrization. Secondly, we propose a first test setup for the acquisition of handwriting traces on paper. The corresponding data is captured from the first and second layer of a five sheets paper stack. Thirdly, we present and discuss our first results of digital image pre-processing based on topographic information representing the pen pressure during the writing process. Our pre-processing approach is focused on segmenting the writing trace in topographical data. For segmentation of the trace we apply different linear filter like mean or binomial as well as non-linear filter like median.
Date of Conference: 03-04 March 2016
Date Added to IEEE Xplore: 09 April 2016
ISBN Information:
Publisher: IEEE
Conference Location: Limassol, Cyprus

References

References is not available for this document.