| Next generation of automatic pattern recognition systems for forensic shoe track applications |
 |
| Researcher |
Adam Kortylewski |
|
|
| Project Partner |
forensity ag |
|
|
| Funding |
CTI-Project 13932.1 PFES-ES |
|
|
| Project Duration |
01.06.2012 - 31.05.2014 |
|
|
|
Footwear impressions are among the most frequently secured types of evidence at crime scenes. For the investigation of crime series they are among the major investigative notes.
In this project, an unsupervised footwear retrieval algorithm was developed which is able to cope with unconstrained noise conditions and is invariant to rotational and translational transformations. A main challenge for the automated impression analysis is the separation of the actual shoe sole information from the structured background noise.
|
|
Typical crime scene shoe print (left) with corresponding reference image (right)
We approach this issue by the analysis of periodic patterns. Given unconstrained noise conditions, the redundancy within periodic patterns makes them the most reliable information source in the image.
|
The impression processing pipeline is built up as follows:
First, a local measure of periodicity is established by fitting a periodic pattern model to the image.
Second, based on the model, the orientation of the image is normalized and the window size for a local Fourier transformation is computed. In this way, distortions of the frequency spectrum through other structures or boundary artefacts are avoided.
Third, the pattern is segmented through robust point-wise classification, making use of the property that the amplitudes of the frequency spectrum are constant for each position in a periodic pattern.
Finally, the similarity between footwear impressions is measured by comparing the Fourier representations of the periodic patterns. Robustness is demonstrated against severe noise distortions as well as rigid transformations on a database with real crime scene marks.
|
