Professional Documents
Culture Documents
Subhadeep Dey
Computer science engineering
SRM institute of Science and
Technology
Chennai,India
Subhadeep.dey30@gmail.com
Abstract—The growing population also leads to a chatbot in the supermarket to speed things up when one
growing consumer base which increases the load and needs to buy something. This chatbot can guide the
resources to cater the needs of the day-by-day shoppers about the location of each product in the store,
increasing consumers. Just as technology is being used its cost, other products similar to it and much more.
to solve all the problems so why not this one. The idea
is to make a more convenient and advanced super
market experience, where there are no cashiers or
lines so that we can shop hassle free. This project
implements Artificial Intelligence and Internet of
Things to automate a supermarket for better
efficiency. A person just needs to walk in after
scanning their QR code through the app with their
unique ID on it. It will record the customer’s presence
in the shop. The shopping carts and baskets have
sensors on them, which can detect the product entering
or being taken out. The products are placed on shelves
which have pressure sensors to detect if any product is
picked up.These shelves are closed and they only open
with your shopping cards. When you reach the
counter, you only have swipe your card again and
money will be deducted from your account according
to your purchase and a receipt will be given. We look
forward to provide this technology to various super
market chains in the country and abroad and help
them implement it with a nominal one time
investment.
can be uploaded to an online cloud service and used for 2. QR Entry- Every customer must scan the QR code from
further analysis and study. Making use of OpenCV can his/her app which will confirm the entry of the customer in
help in eliminating many of the sensors which have been the supermarket and also store the data related to that
prominently used in the previous works done by customer in the database (AWS) with help of the QR scan.
researchers on the same concept. OpenCV will enable 3. Exploring- After the entry point each customer is free to
the system to detect and identify the various faces of the visit any aisle which contains shelves full of different
customers and also products which they are looking to products. Every product will have its unique product code
buy. which is recognisable by OpenCV cascades. These codes
will be right by the camera when the product is being
removed from the shelf and then the algorithm will be able
II. RELATED WORKS to detect which product it is. Another camera will also
A. Chatbot recognise the face of the customer and associate the
products which he or she is holding to their virtual cart. In
There are chatbots available which have been created
a similar way as the customer keeps on taking the products
using Python. Also, some chatbots for supermarket are
already present but they have been implemented using from various shelves These products keep on getting added
Facebook Messenger. This project will feature a chatbot on to the virtual cart of that particular customer. Since all
created using the python library chatterbot which will the products of a similar kind will have the same product
combine both the two by having a Supermarket Chatbot code on them it will be easier to sort the products. For
implemented on Python. instance, if by mistake any customer puts a different
product on the different aisle where it doesn’t belong then
B. Face Recognition a buzzer goes on.
Python has been used to detect faces with the help of 4. Information and Advertising- The chatbot developed using
numpy and OpenCV. Various machine learning algorithms chatterbot library on Python is used to tell the customers
are used in to detect faces using OpenCV which is by far the about the new products that are going to launch or which
most popular library for computer vision. Every object or in have launched recently. More primarily it will work as an
this case face is further divided into smaller bite-sized tasks assist for people who are in a rush and want to know the
known as classifiers. There can be more than 5000 classifiers whereabouts of a particular product.
for a particular face. Checking each classifier to detect a face 5. Exit- When the customer is done shopping for the
can be tedious and time consuming task so to make it products he can simply exit by again scanning the QR code
executable in real-time it uses something called cascades. The
via the installed app and leave. A few moments after
entire face is subdivided into blocks for a quick and rough
leaving the store he/she will receive a message confirming
test. This quick check determines if a detailed test needs to be
carried out or not. the deduction of amount of the total bill of the purchase.
V. SENSORS USED
For sensing the data from the surroundings in the
supermarket we make use of a few sensors. Fig. 1. System architecture
1.Weight sensor- and generates the bill. When the proceed for checkout it
checks item purchase by the customer and generate the final
FSRs (force sensitive resistor) are sensors that allow you bill and the data to the purchase database. The supermarket
to detect physical pressure, squeezing and weight. They are uses the chatbot which is made using the python though
simple to use and low cost. The FSR has 2 layers which are which customer can get the detail about the product and send
separated by a spacer. The more the more the pressure is the data to the Chatbot database. The supermarket also has a
applied, the more of those Active Element dots touch the Smart cart which used to track customer location and product
semiconductor and the resistance go down. The FRS is used inside the cart which used to verify the product which is taken
in rack so as the product is pickup it detects the product and from the rack is kept inside the smart cart. The customer
send mail the main server. move checkout and generate the bill verify the product which
is linked to the purchase check. The purchase database is
connected to the AWS store analytics which share the data
with the store database.
VIII. CONCLUSION
Such an automated system when implemented gives a
better productivity and also saves the customer’s time taken
to shop at the store. It reportedly reduces the time taken to
shop by each customer by 60% which is very low as
compared to the time taken earlier by them.
REFERENCES
The template will number citations consecutively within
brackets [1]. The sentence punctuation follows the bracket
[2]. Refer simply to the reference number, as in [3]—do not
use “Ref. [3]” or “reference [3]” except at the beginning of a
sentence: “Re