Attention Makers

CATEGORY: ENTERTAINMENT

Indian vs Non-Indian face recognition

MAKERS: Krishna , Noor Fathima Khanum COUNTRY: India

This project is to classify the face of a person as to whether he is Indian face or Non Indian face.

The Purpose

This project is to classify the face of a person as to whether it is Indian face or Non Indian face. We can think of it like how we humans sometime make judgement of whether a person as Asian, Indian etc. We use cutting edge technology to build these smart systems. The system is trained with supervised data by showing Indian faces and Non Indian faces and the system is able to learn to classify the faces.

The Technology

In this project we Transfer learning based approach to teach the machine to classify the face. We take a pre-trained Deep face (VGG-face) architecture trained on millions of faces using CNN and extract the higher level features. This network has the capacity to capture the representation of how faces look like. Once we extract the features for Indian and Non Indian faces we use a small neural network for fine-tuning so that the machine can classify the faces.

Additional Details

VGG-face network is built using deep convolutional neural network. After training the network will have the idea of how face image look like and what features can be present in an image.

Home Previous Next

Vote Share Comment