Age and gender prediction using computer vision and convolutional neural network CNN

There are considerable quantities of available data about a person like human face such as age, gender, ethnic background and facial expressions). Humans can easily identify and analyze this information, for example, most people can recognize human characteristics, such as gender, where they can only see their face and tell if they are men or women. In the same context, they can figure out someone's age including determining if this person is still a child or became an adult. Additionally, it is challenging for the computer to build applications that could recognize people's faces and determine their age and gender in which the modern world will depend on many important aspects of our daily life to create a general model that fits all humans. This study focuses on a deep learning solution for predicting age and gender from people's faces. We perform transfer learning and fine tuning from the IMDB-Wiki Dataset using Convolutional Neural Networks (CNN). First, we use the transfer model using three models: the MobileNet V2 model, which is faster to train and is pre-trained with the ImageNet dataset for image classification. The second model, Inception V3 it consists of many convolution and maximum pooling layers. Both models are from Keras application. And we follow the SSR-Net architecture, which is the default for the latest training, with some change