Advanced algorithms for digital medium forensic analysis

Forenzika digitalnih medija pomaže u razotkrivanju zločina kao i spašavanju oštećenih podataka. Kako raste količina prostora za pohranu, tako raste i potreba za unaprijeđenjima u ovim polju. U ovom radu ukratko je opisana digitalna forenzika te forenzika digitalnih medija. Opisani su osnovni problemi koji otežavaju rekonstrukciju datoteka. Prikazane su neke tehnike za slaganje fragmenata i prepoznavanje sadržaja koje su istražene u posljednjih desetak godina. Napravljen je pregled korisnih alata koji mogu pomoći u ovom postupku. Predložena je nova tehnika raspoznavanja datoteka temeljena na statističkim testovima za slučajne nizove. Tehnika je pokazala osrednje rezultate.Digital medium forensic methods help in crime investigations, as well as in practical data recovery scenarios. As the amount of storage space increases, the need for advances in this field grows. In this thesis, digital forencsics was introduced, on which the concept of digital medium forensics was built. Main problems and obstacles during file recovery were desciribed. Some techniques for file reconstruction and content identification developed in the last decade or so were analysed. A list of useful tools that can facilitate in recovery process was outlined. A new technique for statistical file classification based on entropy tests for random sequences was implemented. The technique showed modest results