International Conference on Computer and Knowledge Engineering

Home / 11th International Conference on Computer and Knowledge Engineering

Real-Time Gender Recognition with a Deep Neural Network

Authors :

Samad Azimi Abriz¹ Majid Meghdadi²

1- Zanjan University 2- Zanjan University

Keywords :

Deep learning network, Depthwise separable convolution, Multifold filters, Gender classification;

Abstract :

Nowadays the existence of artificial intelligence (AI) and convolutional networks had led to significant progress in machine vision. Machine vision can automatically perform many tasks that are difficult and arduous and have a high error for humans. One of these difficult tasks is the determination of gender that nowadays has many applications. Using AI and machine vision to determine gender can speed up this process. Deep neural networks have had significant progress in comparison to other common machine learning methods but the number of parameters and calculations is one of the major issues of these networks. In this paper, we have presented a real-time deep neural network model that performs gender recognition faster and with fewer calculations by reducing the model parameters and calculations. The proposed model is a rather light model and a mixture of multifold filters that have been trained and tested on three datasets Wikipedia, Audience, Celeba.