Race Classification through Convolutional Neural Network
Loading...
Date
2017
Journal Title
Journal ISSN
Volume Title
Publisher
University of Management & Technology
Abstract
Neural networks are a powerful technology to classify different images. However, there are impressive number of different types of neural networks that are used in the literature and in industry but we used Convolutional neural network (CNN) to classify the images. The biometric framework can utilize race to distinguish individuals in the world with a precise personality. This research proposes a design to order individuals into two races Asian and Non-Asian that effectively in any face acknowledgment framework can be incorporated. The proposed procedure takes in the assigned essential characteristic of the face, skin color pattern and other secondary feature from training of images in order to effectively classify races. We use CNN to create a system that classifies facial images that are based on a variety of different facial attributes and classify it into two separate classes. We use 3 convolutional layers. We used 3052 training images of 64*64 pixels and we achieve 85% accuracy.
Description
Tayaba Anjum
Keywords
Weber Local Descriptor, Deep Boltzmann Machine, Local Binary Patterns, BS