3. Project Overview

As and when a customer buys something or proceeds for the payment, his/her image will be captured using a webcam and will be stored in an unknown folder. By this time ML models that we have incorporated will be activated and it would predict the Gender, Age, and Sentiments of the customers. That captured image will now be compared with the already existing images of few customers that are stored in a known folder. Now there are 2 scenarios:

• CASE 1: Match Found
  This means the customer is a repeated customer. Then in this case a message will be displayed stating “Match Found” and it would fetch the ID of the image and increment the counter in the database. The captured image will be discarded.

• CASE 2: Match Not Found
  This means the customer is a new customer and has visited the shop for the first time. Then in this case a message will be displayed stating “Match Not Found”. Once this message is received, it will automatically save the image in the known folder as well in the database with a particular ID along with all the details like Age, Gender, Sentiments, etc. while it would delete the image from the unknown folder.

We have made a Server-side website for displaying the data using PUG/HTML/CSS in the front end and Node JS in the backend.