SMART CHECKOUT SYSTEM

Project Title

SMART CHECKOUT SYSTEM

Project Area of Specialization Artificial IntelligenceProject Summary

PROJECT SUMMARY:

Smart Checkout System is less time consuming, more accurate and effortless billing system. This system makes it possible for the people to save their time from long queues at the billing counters of the general stores. Basically, in this system, a camera will be placed at the counter whereas the items will be placed on a particular surface then the camera will detect all the grocery items that are purchased by the customer and will generate automatic billing cart accordingly.

Problem:

In Pakistani general stores and super marts barcode scanning system has been running for a while. Barcode scanner  need a direct line of sight to the barcode to be able to read, they are very labour intensive as each product must be scanned individually, If a barcode is ripped or damaged there is no way to scan the product, barcode scanning is time consuming so people have to wait for their turn in a queue. On the other hand, Smart Checkout System is about to lower down the stress of the people who are standing in a long queue and waiting for their turn at the billing counters of the general stores. our real motive is to neglect the barcode scanning system which is quite time consuming and frustrating process.

Solution:

The solution of this is that, a camera will be placed at the counter whereas the items will be placed on the surface by the customers, this camera will take real time video and detect multiple items from frames After detection, an item will be identified and its price along with its description will be shown and saved in the billing cart and once all the items is identified, the billing cart will be generated with the total amount that a customer has to pay. With that the purchased product will be deducted from the inventory and report will be generated automatically. The system is divided into a series of inter linked modules, each geared towards a special purpose. The first module is of the inventory management module which will manage the database of the respective general store containing all the details and specifications of all the products present in the store and in the warehouses. Then, the product detection module in which model will be trained for the identification of the items after their detection from the belt.After the detection and identification process, a module of cart automation will work in a way that it will generate a bill containing all the items details with their prices also mentioning the total amount of the shopping items which then will be managed by the report management module. All in all, a complete system will be prepared to overcome the time consumed in standing in a long queue for the billing process.

Project Objectives

PROJECT OBJECTIVES:

An objective of our project i.e. Smart Checkout System, is to create an application or system which is able to generate an automated cart by capturing a product and detecting it through a camera to safe the time of customers and effort of the cashier and to increase the usability and provide good user experience. it also briefly describes the problem that was faced by the cashier and the customer while using traditional systems. Concluding it by stating the scope and out of scope of the project.

Our main aim is to recognize and distinguish numerous items in general stores at the checkout counter utilizing picture identification and characterization calculations. The venture will create a automated cart by identifying all the results of clients. This aim can be achieved through steps given below:

• To properly detect the items our camera detects them one by one in order to generate a bill along with their respective prices set in the database. The camera will be placed at the counter where all the items will be scanned. Customers only have to put items on the counter and the rest process will be done by the camera.

•             To train dataset accordingly for the detection of the products at the time of billing. When customer put the items with the possible distance of 10cm, the camera will detect the item with the help trained model and then product details will be extracted from the database.

•  An efficient use of an algorithm YOLOv3 (You Only Look Once) will be implemented to detect the items from real time video for generating the billing cart according to the item prices and quantity.

•  To subtract the item from the cart, like in case any customer wants to minus any item from the cart the screen will show a subtracting and adding sign through which the process can be carried out.

•   To properly manage the inventory through inventory management module. All the items with their respective prices will be managed properly in the database to cater with the products and its identification accurately. If the packaging of an item is not proper, then the system will generate an error so that no worries

Project Implementation Method

Our project is divided into 4 modules:

• Inventory management module
• Product detection module
• Cart generation module
• Report management module

We will be working on basis of modules and each modules have its different implementation.

We will be implementing inventory management module in PYQT5 UI framework through which we will develop User Interfaces of our desktop application and on backend we will be using PYTHON with MYSQL database for transactions of inventories.

Second Product detection module includes detection of products which will be placed on counter so firstly we will gather our data of different products and we will be making our own model using Tensorflow for our detection purpose. We will use YOLOv3 for our object detection task which is highly recommended for real time detection as our requirement is we want to detect products on realtime.

Third Cart Generation module includes the generation of cart in this we will be generating cart on basis of the detected product. We will be using OPENCV in order to interact with our model and extract the details of product that is detected and then it will be added to cart.

Fourth Report generation module will be implemented by using python a complete record will be maintained and reports will be generated accordingly.

Benefits of the Project

Benefits of Smart Checkout System which are as under:

1-Cost of Space

With regards to the cost of space, smart checkout systems counter unquestionably occupy less room than registers. Since self-checkout is commonly quicker than registers, there is regularly one line for every few checkout counters, which saves extra space. However, contrasted with other self-checkout choices, counters are off guard – they are one of the most un-ideal choices with regards to the cost of space.

2-Efficiency

With regards to effectiveness, here are a portion of the contemplations about self-checkout counters:

• Decreased stand by times contrasted with using a cashier lane. Be that as it may, the speed of exchange is fundamentally lower contrasted with a portion of different systems.

• Decreased work costs as one individual from staff can ignore a few self-checkout counters.

• High client commitment: purchasers need to check item standardized identifications and spot items into a packing region, and now and then even gauge items.

• An undeniable detriment of self-checkout, by and large, is shoplifting. A store that has 55 - 60% of exchanges experiencing self-checkout can anticipate that misfortunes should be 31% higher. Nonetheless, with diminished expenses and continually improving security, this rate is diminishing after some time.

3-Customer experience

Speed of transaction perceived control, reliability, usability and delight are the most in?uential factors prompting consumer loyalty. Considering, this type of checkout is at a favorable position contrasted with sales registers, however, it is behind different frameworks recorded underneath. As self-administration checkout counters have been around for some time, most clients are very used to this experience. Despite the fact that self-checkout counters do advance the utilization of room, effectiveness, and consumer loyalty, their usability can be improved. As per an examination, 74% of clients state that trouble in entering merchandise and successive supersedes were their greatest worries with the innovation. 59% of buyers said that they would be bound to utilize self-checkout if the innovation was improved.

Technical Details of Final Deliverable

Technical Details of Final Deliverables:

Our final deliverable would be a complete smart chackout system in which product will be detected thorugh IP camera and on behalf of each product details will be extracted from database and cart would be generated accordingly and then reciept will be printed.

IP camera will send real time video to the system and system will divide the video in frames and on each frames detection are performed.

Deployement Strategy and working in Production enviroment:

Our desktop application will have admin and retailer side interfaces will be deployed at supermarket with an IP camera installed on the counter with a certain angle  that will capture products. Our Backend API will be deployed at cloud server which will be interacting with our model that is also deployed on server.

Desktop application will recieve video from IP camera and then frames are extracted and each frame will be sent to API for detection of product and final response will be a detected product with its price and other details which willbe added to cart.In this way application will work in working enviroment

Support Document:

We will be providing complete book contains all the details of the project and its implementation. This book will contain the algorithm we used all the literature work and all the uml diagrams with its explanation. We will be also providing a shirt tutorial of how to use our application

Final Deliverable of the Project HW/SW integrated systemCore Industry OthersOther IndustriesCore Technology Artificial Intelligence(AI)Other TechnologiesSustainable Development Goals Decent Work and Economic Growth, Industry, Innovation and InfrastructureRequired Resources

Item Name	Type	No. of Units	Per Unit Cost (in Rs)	Total (in Rs)
			Total in (Rs)	10000
IP camera	Equipment	1	10000	10000