DEVELOPMENT OF NEW SEGMENTATION ALGORITHM FOR UNCONSTRAINED EAR’S IMAGE
|
|
Human identification biometrics systems may originate from real-life criminal and forensic applications. Various methods, such as finger printing and face recognition, already proved to be very efficient in computer vision based human recognition systems. Nevertheless, forensic biometrics is still need to be improved. Numerous studies are attempt to propose novel and innovative solutions for new biometrics modalities such as human scent recognition, EEG biometrics, finger tips tissue recognition or ears recognition. Interestingly, ears biometrics has some advantages. Unlike the face that will change considerably aging, the ear is one of the most stable human anatomical feature. Furthermore, ears are captured without any contact to special device. The study in ears biometrics is still on progress. Recent works shows recognition rates on clear and close-up ears image can achieved promising result. Unfortunately, only few explored ears recognition that captured on different angle. This issue gives a motivation to this research to develop techniques or algorithms that can achieved promising recognition rates. The ideas of this research are divided into two stages. First, the ears image is enhanced during preprocessing step. This step is proposed to outcome best ears image for further detection and segmentation. Second, the efficient ears segmentation method will be explored.
|