Age, Gender & Emotion Detection Using CNN For Retail Analytics Customer Profiling

Abstract

These days computerized age & gender detection is a completely exciting research topic. For My this studies, when someone enters a retail shop, his or her photo is captured via the security camera. That image will be despatched as input to a machine learning model & by using those ML models age, gender & emotion would be predicted. Then the predicted facts will be stored in a database, for consumer profiling. The intention of this study is to research the usage of image detection for making better customer profiling for retail shops, & to make extra profit. Such as if it’s far discovered in a store, most people come between the age group 20-50 & what’s the most coming gender, then they can plan a shop offer for those products in which that specific age group or gender’s people are interested, however, it will additionally assist to promote marketing, particularly for the age group of human beings. Then again when a purchaser enters a retail store & that person’s emotion is diagnosed, then through knowledge of shoppers’ sentiments & evaluation of their emotional responses, outlets can keep better customer service & assure first-class service towards each of the customers. A 2D Convolutional Neural Network (CNN) version is constructed to predict people’s age, gender & emotion. The version accurately carried out 82.5% for age, 89.5% for gender & 97.3% for emotion.