Seminar Details

Seminar Title
::
Development of Ear Biometric Systems with Forensic Validation
Seminar Type
::
Defence Seminar
Department
::
Computer Science and Engineering
Speaker Type
::
Student
Speaker Name
::
DEBBROTA PAUL CHOWDHURY ( RollNo : 516CS1003)
Date  &  Time
::
13 Jun 2021  03:00 PM
Venue
::
Online, Google meet link: meet.google.com/emt-grvk-gtu
Contact
::
Prof. Sambit Bakshi
Abstract
::
Choosing an appropriate biometric trait is a critical job for robust and meticulous performance of a biometric authentication system. Generally, it is observed that face is the most feature dense region in human body. Face, iris, lip, and ear are some of the examples of biometric traits, that are occupied within the facial region. Automatically authentication of individual is an ever growing field of research. Several research over the course has been done since last decade and many researchers found ear as an important biometric trait for automatic recognition. Ear biometric has several advantages over other biometrics. Unlike face biometric it is not affected by facial expressions, makeups and specs. There is an aging effect on face as well as fingerprint but not in ear. Overall, we can say that, ear biometric is ideal for active as well as passive recognition due to the robustness of features and image acquisition from a distance. The defence seminar on the doctoral thesis mainly concentrate on two application areas in ear biometrics: (a) Ear recognition, and (b) Forensic validation of ear biometric system. Handcrafted as well as deep features are extracted from ear images to recognize a person using ear biometric. Then forensic validation of ear recognition system is proposed, and finally forensic validation is performed with reduced feature vector size.