.png){kind=icon}
Return
Ear-based authentication using information sets and information modelling
.png){kind=icon}
PRE
AI
DOI：10.1007/s00500-021-05858-3.png)
.png){kind=icon}
Original request for help.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
As the ear biometric needed for forensic analysis and surveillance is acquired under the unconstrained environment, it suffers from the problems of illumination, pose and occlusion. To mitigate these problems, representation of both possibilistic certainty and uncertainty in the gray levels termed as the information source values of ear biometric is attempted at using the Hanman-Anirban entropy (HAE) function. An adaptive form of this entropy function helps derive Hanman transform that represents the higher level possibilistic certainty. The sum of information values constituting the information set gives the possibilistic certainty/uncertainty. The information values are modified to generate information set features such as sigmoid, energy, effective information source and effective information. Hanman filter is developed to modify the frequency content of the information values. The information rules are framed to facilitate the unsupervised learning and a new classifier called Weighted Hanman Classifier (WHC) is developed using t-normed error vectors and HAE function.WHC outperforms the other classifiers on the proposed feature types for the ear-based authentication under both constrained and unconstrained conditions.
Keywords:
Ear biometric
Information sets
Hanman transform
Hanman filter
Weighted Hanman Classifier
Journal
IF:
2.5
Papers:
1.0W
Citations:
2.1W
Organization
Cited Papers
.png){kind=icon}
ShareSaveShareSaveShareSaveShareSave
View more