A novel approach for face detection using artificial neural network

Author

Quraishi, Md Iqbal ; Das, Goutam ; Das, Aruneema ; Dey, Prasenjit ; Tasneem, A.

Author_Institution

Dept. of Inf. Technol., Coll. Kalyani, Kalyani, India

fYear

2013

fDate

1-2 March 2013

Firstpage

179

Lastpage

184

Abstract

In recent time face detection is of utmost importance because for its various applications. Several approaches have been implemented to date. This paper aims towards an effort to represent a novel approach for human face recognition. The proposed system consists merging both frequency and spatial domain techniques. The proposed system selects the Region of Interest on which Ripplet Transformation is to be applied after power law transformation to calculate Standard Deviation (SD) and Mean as features. At later stage, Feed Forward Back Propagation Neural Network (FFBPNN) is used for classification and recognition purpose. The approach is tested with non face images to show its effectiveness which is around 91.67%.

Keywords

backpropagation; face recognition; feedforward neural nets; frequency-domain analysis; image classification; transforms; FFBPNN; SD calculation; artificial neural network; face detection; feedforward backpropagation neural network; frequency domain techniques; human face recognition; mean calculation; nonface images; power law transformation; region of interest; ripplet transformation; spatial domain techniques; standard deviation calculation; Biological neural networks; Face; Face detection; Face recognition; Feature extraction; Transforms; Mean; Power Law Transform; Ripplet Transform; Standard Deviation; classification; face detection; face recognition; feature extraction; feed forward back propagation artificial neural network; interpolation;

fLanguage

English

Publisher

ieee

Conference_Titel

Intelligent Systems and Signal Processing (ISSP), 2013 International Conference on

Conference_Location

Gujarat

Print_ISBN

978-1-4799-0316-0

Type

conf

DOI

10.1109/ISSP.2013.6526898

Filename

6526898