Gender Prediction

Introduction

The automatic estimation of gender from face images has recently attracted attention due to its wide adoption in various facial analysis applications, especially since the rise of social media platforms.

Gender estimation from facial images plays a large role in many applications, like marketing intelligence. Among these applications, the Android application, AgeBot, finds multiple faces in a picture and estimates the age and gender of each localized face.

Objective

In this lesson, we’ll use a pretrained, deep learning gender detector model that was initially implemented and trained by Gil Levi and Tal Hassner to predict gender using facial images. For additional details, please refer to their publication, released in 2015, “Age and Gender Classification Using Convolutional Neural Networks.”

These two researchers have already published their CNN gender detector model as a Caffe model. Inherently, the OpenCV library supports loading and exploiting neural network models of the type Caffe.