You are on page 1of 23

SEAT STATUS

DETECTION
“EARLY DETECTION IS A BETTER
CHOICE TO SAVE TIME”
INTRODUCTION
• Finding an empty chair is one of the most subtle and common problems in
everyday life.

• People face Challenges in finding table at restaurants.

• Groups or families face similar issues when trying to find suitable seat sets.

• Huge seating capacity leads to issues with new customers searching for empty
table.
SOLUTION
• we have developed a detection system called “Seat status Detection”.

• The system can detect the empty table in real-time.

• The user can easily see the number of tables empty or filled from anywhere
through our application.

• The user can book the table before arriving through application.
IMPLEMENTATION
FRONT-END
• We have made the following two mobile applications with the help of flutter,
which is Google's portable UI toolkit for crafting beautiful, natively compiled
applications for mobile.

• Admin Application.
• User Application.
ADMIN APPLICATION
• Admin app may used by the restaurants manager.

• The admin have the authority to do the following actions in the app.

1. Admin can view and also block the tables if some issue occurs.

2. Admin can view customers in penalty and can remove them from penalty.
ADMIN APPLICATION

3. Admin can also add and remove the food items and deals.

4. Admin can check the customer queries.

5. Admin can add the discounts on foods for users.


USER APPLICATION

• User app may used by the customers.

• The users can use the app from anywhere.

• The user have the authority to do the following actions in the app.
USER APPLICATION

1. The users can easily view the number of tables available or not in the restaurant from
any where.

2. The users can book the table before arriving, the table will be reserved for the user.

3. The booked table will come out from penalty if the user doesn’t arrives.
USER APPLICATION

4. The user can view the popular food items, deals of the restaurant.

5. The user can check the discounted food items.

6. The user have the option to edit the profile.

7. The user have the option to rate the restaurant and see the location of restaurant on map.
DATABASE
• We used firebase, which is a popular database.

• The reason of using firebase was because it is used for storing and retrieving data
in real-time.

• It provides a cloud-based NoSQL database solution, which means data is stored in


the form of key-value pairs and can be easily accessed and manipulated in real-
time.
• We created following documents in our database using firebase.

o User information: Details of customers and restaurant employees like name, email,
contact etc. who interact with the seat detection system.

o Table Dataset: we stored table dataset taken by us which we get from back-end.

o Occupancy data: Records of the seating status (occupied or unoccupied) of each table in
the restaurant.
o System logs: A record of all actions performed by the system, such as table status
updates and alerts.

o Contact us: The feedback and queries entered into the "Contact Us" section is stored in
the database.

o Products: The food details, id, name, price etc. are stored in database.

o Deals: All food details of the restaurant is stored in database.


BACK-END
• Deep learning is a subset of machine learning that is based on artificial neural networks
(ANN’s) with multiple layers, also known as deep neural networks (DNN’s).These neural
networks are the structure and function of the brain, and they are specially designed for
learn large amounts of data in an unsupervised and semi-supervised manner.
• Deep learning models can automatically learn features from the data and their tasks such
as image recognition, speech recognition and natural language processing.
• There are three types of architectures that are widely used in deep learning.
1) Feed forward neural network (FNN’s)
2) Convolutional neural network (CNN’s)
3) Recurrent neural network (RNN’s)
• is a simplest type of ANN with a linear flow of
Feed forward neural network:

information such as image classification, speech recognition, and natural


language processing.
• Convulotional neural network: is a special designed for image and video recognition.CNN’s can
automatically learn features from the images.

• Recurrent neural network: are a type of sequential data such as time series and natural processing
language.

• Difference between Machi9ne Learning and deep learning:


Machine Learning Deep Learning
Works on small amount of dataset Works on large amount of dataset
Dependent on low end machine Dependent on high end machine
Take less time to train Take longer time to tain
YoloV5:
• You only look once (yolo) this is real time object detection system.Yolov5
processes images using a single neural network, and then divides into different
parts.
• YoloV5n:This is smallest and fastest model. It also called the nano model and
finds application in mobile devices owing its size.
• 800 to 1000 images has been used in the dataset.

• We used openCV, to perform image and video processing task.

• We used OpenCV with Python to draw square boxes around table in video, and
label these boxes with the class of “filled” and “empty”.
• We captured the frame with the help of camera using openCV.

• The PyTorch framework is being used to extract YOLOv5 data/files and integrate
it into the project for further processing and analysis.

• After counting the boxes, it is added to the database.

You might also like