Determination of Age, Gender, Dress Color and Type of a Person by Convolutional Neural Network (CNN)

Abstract

Artificial Intelligence is a soft computing technique that mimics the human intelligence which interprets and percepts based on the inputs given to it, for the output generation or prediction. The input image is processed by the neural network along with the help of the algorithms of Machine Learning making it advantageous to process inputs with reduced use of hardware and the complexities of implementation irrespective of number of models are needed to be developed and deployed. Determining a person’s age, gender and color and pattern of the dress is done by the help of the convolutional neural network and Computer Vision for video analytics such that they can be determined individually and in a combined manner by it at a good value of accuracy in training and validation. The existing works could determine only one but not with the combination of the parameters, so pre-trained models like the Caffe, minivggnet models are used to determine them along with the use of openCV. The coding is done in python language, wherein packages like tensorflow, keras, scikit and openCV are used. Opencv is used for the video analytics such it can be used for the trend analysis or the surveillance purposes.