“Gender Recognition and Age Estimator Using Deep Learning Techniques”

EasyChair Preprint 8070

Abstract

Gender may be a central feature of our personality still. In our social life it's also a significant element. Computing age predictions are often utilized in many fields, like smart human-machine interface growth, health, cosmetics, electronic commerce etc. The prediction of people's sex and age from their facial images is an on-going and active problem of research. The researchers suggested variety of methods to resolve this problem, but the factors and actual performance are still inadequate.
Age and gender that are the two key facial attributes play a foundational role in social interactions, making age and gender estimation from one face image an important task in intelligent applications, like access control, human-computer interaction, enforcement, marketing intelligence and visual surveillance. The fundamental aim of this paper is to develop an algorithm that estimates age and gender of an individual correctly. One of the most widely used techniques is haar cascade. In this paper we propose a model which can predict the gender of a person with the assistance of Haar Cascade. The model trained the classifier with different male and female images as positive and negative images. Different facial features are extracted. With the assistance of Haar Cascade classifier will determine whether the input image is male or female. We made use of Deep- Convolution neural network. It works efficiently even with limited data. For the age approximation task, the paper makes use of caffedeep learning framework. Caffe provides expressive architecture, extensible code. Caffe can process over 60M photos per day. This makes it one of the fastest convent implementation available.

Keyphrases: Age classification, Caffe deep learning framework, Convolution Neural Network, Gender Recognition, Haar Cascade

@booklet{EasyChair:8070,
  author    = {Komal Kamble and Rashmi Ashtagi},
  title     = {“Gender Recognition and Age Estimator Using Deep Learning Techniques”},
  howpublished = {EasyChair Preprint 8070},
  year      = {EasyChair, 2022}}