Year of Graduation
Retail Video Analytics in Age and Gender Classification Problem for Mobile Platforms
In this paper we examine the age and gender video-based recognition problem using deep convolutional neural networks for contextual advertising applications. The literature review of image classification methods is presented. The possibility of implementing the solution aggregation to improve the accuracy of identification is considered. An experimental study has been conducted to determine gender and age from data sets: Kinect, IJB-a, Indian Movie, EmotiW2018. Also in this paper several models of convolutional neural networks are compared, including "light" models, created for implementation on devices with limited resources. As a result, a prototype of a mobile application on the Android platform has been developed taking into account the most accurate convolution neural network model and classifier fusion algorithm.