Hand-based biometric recognition system for mobile devices

• Main Author: Serrano, Luís Miguel dos Santos
• Publication Date: 2011
• Format: Master thesis
• Language: por
• Source: Repositórios Científicos de Acesso Aberto de Portugal (RCAAP)
• Download full: http://hdl.handle.net/10071/8235
• topic: Android; Binary templates; Cryptographic hash function; Error correcting code (ECC); Hand-based; Low-density parity-check (LDPC); Mobile devices; Orthogonal Line Ordinal Features (OLOF); Palmprint; Secure biometric recognition system; Templates binários; Função de hash criptográfica; Código corrector de erros (ECC); Código Low-density parity-check (LDPC); Dispositivos móveis; Palma da mão; Sistema de reconhecimento biométrico seguro

The usage of mobile devices such as smartphones has become a daily habit and need, as a way of managing and accessing personal data on the go. This created the emergent need to secure that data, which may contain not only contacts and messages, but also online accounts, files, bank account information and other applications or services which should be protected in case of device loss or theft. Biometrics have been an increasingly adopted solution, allowing individual recognition based on biometric traits (e.g., palmprint, voice, face, iris, signature), in a more practical and secure way, since biometrics cannot be forgotten or lost and are intrinsically associated with each individual. In this dissertation, a secure hand-based biometric recognition system is proposed and implemented for the Android 2.1 mobile platform. The system makes use of the camera present in Android devices (typically with a resolution of 5 Megapixel or higher) and performs a pre-segmentation of the captured images in the device to help the user optimize the data acquisition result. The data is then transferred over a secure connection to a server which performs further image analysis and feature extraction on the palmprint, using the Orthogonal Line Ordinal Features (OLOF) technique. A system-specific template composed of the extracted features is then securely stored in a database, with the help of a cryptographic hash function and an error correcting code (ECC).