Integration of Non- Decimation and PCA in Wavelet Based Multi-Band Ear Recognition

Ear recognition has gained prominence as a reliable biometric modality due to its unique and stable features. This paper gives an approach for ear recognition using a nondecimated wavelet transform in a multiband framework, coupled with Principal Component Analysis (PCA) for dimensionality reduction...

Full description

Saved in:

Bibliographic Details
Published in:	2024 15th International Conference on Computing Communication and Networking Technologies (ICCCNT) pp. 1 - 6
Main Authors:	Vemuri, Ratna Kumari, Anandamai, Vemu S V S S, Ganesh, Vatturi Prasanna, Jyothi, Nalamala Sai, Anil, Palaparthi
Format:	Conference Proceeding
Language:	English
Published:	IEEE 24-06-2024
Subjects:	biometrics Ear Faces Feature extraction Image segmentation multi band frame work non decimation Principal component analysis principal component analysis(PCA) Reliability sub bands Vectors Wavelet analysis Wavelet coefficients wavelet transforms
Online Access:	Get full text
Tags:	Add Tag No Tags, Be the first to tag this record!

Description
Summary:	Ear recognition has gained prominence as a reliable biometric modality due to its unique and stable features. This paper gives an approach for ear recognition using a nondecimated wavelet transform in a multiband framework, coupled with Principal Component Analysis (PCA) for dimensionality reduction. The suggested technique for 2D-WMBPCA involves dividing the input image into wavelet subbands using a 2 dimensional wavelet transform that retains all of the wavelet coefficients. This wavelet transform avoids down-sampling, preserving more information and allowing for better feature extraction. The number of frames in each resulting sub band is divided based on the coefficient values. Equal size can be used to calculate the limits of multiple frame generation. The multiband wavelet coefficients dimensionality is then decreased by applying PCA. It helps in selecting the most discriminative features and mitigates the curse of dimensionality, contributing to improved recognition accuracy and faster computation.
ISSN:	2473-7674
DOI:	10.1109/ICCCNT61001.2024.10724982