August 2009 (2)

This month I suddenly had to go to a court case in Amsterdam to testify. I had to wait from 2 to 6 pm in the court and witnessed a battle of experts. After that I could testify in the evening, and it gave me also strengthened me in the opinion that good validation, and testing different hypothesis is important in forensic science, together with clear interpretation and statistical evaluation in an understandable way for the court is needed. This was also one of the main issues within the NAS report of Strengthening Forensic Science in the United States.

Currently I am at the DFWRS- conference, and just gave a keynote talk on Opportunities and Challenges in Digital and Multimedia Sciences. It is an excellent conference where also the issue of having good validation based on ground thruth databases is discussed, and the issue of using software for analyzing forensic evidence with minimal standards is reproducible and reliable. The issue of the paradox of the zero-skill tools which are most optimal for wide use within police was also discussed. I also discussed our PRNU-tool with the new version where more batch processing is possible. We need also large databases for open source. Also the camera databases would be excellent to have for verification and validity of camera comparison.