Real time promotion and marketing based on customer diversity

Project Title

Real time promotion and marketing based on customer diversity

Project Area of Specialization Artificial IntelligenceProject Summary

A web based application which is designed for retailng industry that will allow marketers to have a complete record of their customer diversity. It will enable them to promote retailing brands according to the cusotmers' age and gender present at that period of time in malls.

Project Objectives

1) To design a stand-alone software for super stores to target their audience.

2) To classify gender and age group of customers on the basis of facial features.

3) To analyze and identify the peak periods.

4) To help retailers to get a clear profile of their customer and adapt signage to attract and engage more customers.

Project Implementation Method

We are using Sequential Fully Convolutional Neural Netowrk for classification of age and gender on the basis of facial features of customers' entering/leaving in malls/retailing stores. Training data in the form of human faces of 516785 is being used for classificaton of age and gender. Live feed for testing is being taken by camera for real time classification. Statistical segmented data will be shown to dashboard in the form of graphs. 

Benefits of the Project

According to research* Pakistan has the most flourishing retailing industry.  **The retail sector constitutes 33% of the overall service sector in Pakistan and contributes around 18% to its total GDP. The country’s retail industry is valued at USD 42 billion. To help our marketers to flourish further we have devised a solution which will help retailing
industry to boost their sellings by promoting different retaling brands based on customer diversity.

Links for research mentioned above 

*http://dunyanews.tv/en/Business/407650-

**https://www.emergingpakistan.gov.pk/opportunities/punjab/retail/

Technical Details of Final Deliverable

A stand-alone web application that takes live feed from  camera and analyze statistical information about gender and age group segmentation of customers and display final results on the dash board.

We are using Python(Keras, dlib and tensorflow) for classification and boot strap and Google visualization APIs for dashborad development. Our system will be connected to database of SQL Server using PHP. 

Final Deliverable of the Project HW/SW integrated systemType of Industry Others Technologies Artificial Intelligence(AI)Sustainable Development Goals Industry, Innovation and InfrastructureRequired Resources

Item Name	Type	No. of Units	Per Unit Cost (in Rs)	Total (in Rs)
			Total in (Rs)	54600
Movidius Neural Compute Stick	Equipment	2	15000	30000
Respberry pi zero V1.3	Equipment	2	2900	5800
Respberry pi zero 1.3 camera	Equipment	2	900	1800
3-D casing and packing	Miscellaneous	1	1500	1500
Cloud Storage and computation	Equipment	1	7500	7500
Contigency/Overhead	Miscellaneous	1	3000	3000
Subscription Charges	Miscellaneous	1	3000	3000
Operations+Market	Miscellaneous	1	2000	2000