Burglaries, especially in private homes, are because of their large number very unsettling for significant parts of the population and require from the police a high investment of resources for their clearing.
Tool marks resulting from the break-in tools significantly support the investigation of such offenses and are crucial and explicit evidence in the following court cases. So far, the interpretation of tool marks is very labour- and time-intensive and requires skilled personnel for the comparison of different traces with one another, as well as with the regarding burglary tools. At present, the comparison is usually limited to only confiscated tools and some few traces of current crime scenes, which renders a detection of e.g. burglary series throughout Austria impossible. With the possibility of a comprehensive tool mark comparison, such strings of events could be unerringly linked to one another.
Above of that this activity is very stressful for the deployed staff and it is difficult to keep up the task for an extended period of time without the risk of error. Currently, there is no retrieval systems specialised on tool mark comparison available on the market.
For this bottleneck FORMS provides fast, innovative methods for (semi-)automatic search and retrieval of similar tool marks in criminal offenses. It enables the automated comparison of both mark and tools, as well as of one track with another track, such as in the case of crime series. The data from unsolved crimes and new data are merged in a database structure and made available for the query so that in the future it will be possible to compare tool traces on a large scale. With this solution the workforce in the forensic examination will be relieved to manually compare only a limited selection of very similar tracks pre-selected by the FORMS software.
Since different tool marks from one tools may have complex dissimilarities due to various factors such as wear, approach angle, etc., FORMS applies computational methods for image (texture) comparison, e.g. features based on Fourier Transform, Wavelets, Gabor Filters, Local Binary Patterns, etc., in combination with Machine Learning methods.
With the help of FORMS, the potential of tool marks can be significantly more effectively exploited for the forensic investigation of crimes and will provide an important contribution to the subjective feeling of safety for the citizens.
Technische Universität Wien
Institut für Rechnergestützte Automation
Univ.Prof. DI Dr. Robert Sablatnig