Local Line Binary Pattern Vein Extraction from Dorsal Hand Images for Intruder Detection

Personal identification techniques could be found in a wide range of systems that process biometric data. Nowadays, fingerprint, voice or iris traits are employed on a large scale for biometric identification. In recent years, the hand vein structure models attracted the researchers' attention and is considered to be much safer compared to other `classical' biometric features. The paper proposes a method of intruder detection using dorsal hand vein pattern images, from a database with 102 users (classes). The vein extraction has been done using the Local Line Binary Pattern (LLBP) technique. The intruder detection module uses the Dice coefficient for both intraclass and interclass similarities computation.