Automated Shoplifting Detection System

Shoplifting is one of the most common cause for smaller commercial business to fail in today?s market and become most common type of theft in the United States [1]. In the past, three most common methods-electronic labeling, store restructuring and placement of uniformed guard were used to prevent s

Project Title

Automated Shoplifting Detection System

Project Area of Specialization

Artificial Intelligence

Project Summary

Shoplifting is one of the most common cause for smaller commercial business to fail in today’s market and become most common type of theft in the United States [1]. In the past, three most common methods-electronic labeling, store restructuring and placement of uniformed guard were used to prevent shoplifting [2]. None of these techniques is sufficient for large scale implementation of shoplifting detection

, There is no state of the art method for automatic monitoring and detection of shoplifting activity. Previous methods were introduced only to minimize shoplifting activity by using convex mirror and store redesign but, cannot detect shoplifting directly So, we need to make an automated system that detect shoplifting from live streaming of videos.

Now-a-days, there is a need of developing a shoplifting activity detection and recognition system that our retailers can use for reducing the shoplifting activities by monitoring these activities through an automated system that provides them efficient and reliable services

Project Objectives

Now-a-days, there is a need of developing a shoplifting activity detection and recognition system that our retailers can use for reducing the shoplifting activities by monitoring these activities through an automated system that provides them efficient and reliable services

.

Shoplifting is distributed over a large area in wholesale establishment but, still exact information on its extent is not available hence it is not clear whether the stealing is committed by the customers, staff or supplier [2]. There is a need of system that detect and recognize these criminal activities and alarm the security team about these activities so that time for continuously monitoring these activities reduces.

Shoplifting is one of the most common cause for smaller commercial business to fail in today’s market and become most common type of theft in Pakistan. In the past, three most common methods-electronic labeling, store restructuring and placement of uniformed guard were used to prevent shoplifting. None of these techniques measurement is sufficient for large scale implementation and have limitations. To enable quick response for such criminal activities, efficient automated shoplifting detection system is needed.

To develop Automated Shoplifting Detection System, it’s a challenging task to implement video processing using deep learning techniques and now a day, it’s a research issue to develop fast and efficient algorithms to get high accuracy by using video processing technique to manage large datasets.

A benchmark dataset for research community.

Project Implementation Method

we will use deep learning techniques for this project. we will extract features of the frame with CNN and train a LSTM model on these features.

Both iterative and incremental development model are preferred for this project because we will be testing our project several times during the development to achieve the maximum accuracy. From four modules, one module is implemented and tested for accuracy and so on with other modules.

Programming language Python will be used in our project. Our primary host platform is Anaconda for LINUX and secondary for WINDOWS. From training to production, Anaconda is an AI/ML platform that allows organizations to develop, govern, and automate AI/ML and data science with the help of Python language.

For one system with a camera the budget will be:

• 30-40k for system (approx.)
• 60-70k for GPU GTX 1080Ti/1050Ti (8/11GB) approx.

Benefits of the Project

Our system will detect and recognize shoplifting activities in malls. Our system will detect the shoplifting in the following cases.

1. When someone tries to put goods in their pocket.
2. When woman tries to put goods in their bags.
3. When students put goods in their college bags.

our major task will be:

• Shoplifting detection and recognition.
• record their frames.
• generate alerts when shoplifting occurs.
• maintain a record of these shoplifting activities that can be used for security purposes.

So that the person who will be monitoring the whole system act against them. Our main target will be:

• reduce men force.
• Increased accuracy

Technical Details of Final Deliverable

final deliverable will be a complete automated system . it will be a desktop application that can be deployed on any mall with cameras.

Final Deliverable of the Project

Software System

Core Industry

Security

Other Industries

Core Technology

Artificial Intelligence(AI)

Other Technologies

Sustainable Development Goals

Peace and Justice Strong Institutions

Required Resources

If you need this project, please contact me on contact@adikhanofficial.com