Ink-deposition Analysis Using Temporal (online) Data - Inria - Institut national de recherche en sciences et technologies du numérique Access content directly
Conference Papers Year : 2006

Ink-deposition Analysis Using Temporal (online) Data

Abstract

This paper focuses on a new computational method for discovering and evaluating ink-trace characteristics related to the writing process. Typically, these ink-trace characteristics are inner ink-trace morphology, trace width and / or ink-intensity variations, but also specific stroke phenomena, such as ink drops, feathering and striations. The aim of the evaluation is to provide objective and reproducible analysis results for stroke morphologies in order to detect skilled forgeries and technical copies. The influences of different kinds of writing material, like the ink type used, is taken into account. By means of recorded, temporal (online) data, segmented ink traces are sampled equidistantly, and local ink-trace characteristics are encoded in one feature vector per sample record. These data establish a sequence which faithfully reflects the spatial distribution of ink-trace characteristics and solves problems of methods previously available. Dynamic Time Warping is implemented for the verification of two feature-vector sequences. The proposed method works towards (1) detailed studies of ink-deposition processes, (2) objective testing procedures in forensic practice, and (3) the advancement of skilled forgery detection for automatic bank-check processing.
Fichier principal
Vignette du fichier
cr1060192277313.pdf (691.28 Ko) Télécharger le fichier
Loading...

Dates and versions

inria-00104286 , version 1 (06-10-2006)

Identifiers

  • HAL Id : inria-00104286 , version 1

Cite

Katrin Franke, Steffen Rose. Ink-deposition Analysis Using Temporal (online) Data. Tenth International Workshop on Frontiers in Handwriting Recognition, Université de Rennes 1, Oct 2006, La Baule (France). ⟨inria-00104286⟩

Collections

IWFHR10
191 View
368 Download

Share

Gmail Facebook Twitter LinkedIn More