Introducing TRIGRAPH - Trimodal writer identification

With this presentation, we want to introduce the recently started TRIGRAPH project. The aim of this project is to develop techniques that can be used for forensic writer identification, using recent advances in pattern recognition and image processing, new insights in automatically derived handwriting features, user interface development, and innovations in forensic writer identification systems. The challenge in TRIGRAPH is to integrate these recent developments into a test-bed system, with the goal to improve the writer identification systems available today. A particularly well-established method in forensic document examination exploits allographic information, i.e. particular shapes in characters. A technique that can be used to compare two allographs is Dynamic Time Warping (DTW). We believe that DTW performs this comparison so that the results are visually convincing to humans. To test this claim, we conducted an experiment in which non-expert subjects were asked to judge the quality of the DTW-results, in comparison to the results of another system. We found that one particular implementation of DTW was significantly rated as the best system. The next step is to evaluate whether or not forensic handwriting experts will also judge the results of DTW to be more convincing than the results of other techniques.