Project: Advanced Customer Behavior Analysis - Tracking of Customers and Goods in Retail Shops based on Intelligent Video Surveillance
We are facing a revolutionary change in the field of video surveillance systems. The traditional analogue CCTV systems are replaced by digital, IP based solutions almost everywhere, and real time video processing is also becoming feasible and cost effective. With the application of machine intelligence, the traditional storage and retrieval based surveillance systems are extended with new automatic functions to help the security personnel to recognize effectively different suspicious and security-relevant events and situations._x000D__x000D_In parallel with the spreading of intelligent functions for the purpose of security, several other areas have emerged, where such intelligent video techniques can provide significant added value. In this respect one of the most booming areas seems to be the market analysis in retail industry. As opposed to security, the return of investment in such cases is easily demonstrable, thus it provides a much higher business potential for the already existing technologies. Intelligent solutions can increase profitability and gain advantage over competitors in retail by providing yet unreachable insights into customers' behavior and the effectiveness of some in-store marketing activities. So, market researchers expect the retail segment within the intelligent video surveillance market to grow exponentially within the next five years._x000D__x000D_Video systems can serve different purposes in a retail shop:_x000D_- The first applications included the support for different LOSS PREVENTION investigations in order to reduce inventory shrinkage by the detection of relevant events. This was initially achieved by customizing the existing security oriented products._x000D_- CUSTOMER EXPERIENCE can be improved by further analyses, e.g. by assessing the queuing times, and supporting a decision to open a new cash register in case this time exceeds a certain level. _x000D_- Finally, by more complex and sophisticated intelligent functions, it also became feasible to help retailers make better BUSINESS DECISIONS and increase sales and profitability by analyzing customer behavior._x000D__x000D_Goal of the CUBEA project is to focus on this latter business opportunity, and apply video analytic techniques to track persons and goods in retail shops. Based on the collected information the resulting system will analyze and predict the behavior of customers. Tracking persons up to the cashiers along with their purchases enables the analysis of the correlation between the content of their basket and the path they walked through in the shop. This will help to optimize the placement of goods and advertisements, and as a result it will maximize sales income._x000D__x000D_The system will be able to provide automatic and instant feedback on certain schemes of display locations and placement of goods on the shelves. This way it will replace the usually anecdotal evidence regarding the formed practices, and will replace subjective assessment with objective, statistical evaluation accomplished automatically based on the surveillance camera images. Besides optimizing display and product distribution effectiveness, additional benefits include the possibility to charge premium prices for proven prime display locations, calculate product conversion rates or identifying and profiling some typical consumer groups._x000D__x000D_The operation of the system is based on the automatic detection of certain events and situations by the video analytic subsystems, including the tracking of people, goods and shopping carts. This information is then correlated with the list of purchased products retrieved from the cash register (POS). The products are placed along the routes of the shopping carts and customers, through which the picking order and the shelf-visiting order is deduced. Different correlations between the collected pieces of information will be revealed by geospatial data mining, resulting in both an in-depth analyses and in executive overviews directly supporting business decisions. _x000D__x000D_Some examples of the possible conclusions the system can draw are for example: "The probability for buying both washing powder and then fabric softener depends on the order of purchase: 70% / 30% respectively", or "When buying washing powder, 60% of people go to buy softener; but when buying softener, only 20% of people go on to buy powder". Based on the deduced rules, the CUBEA system will be able to simulate and predict the effects of the reorganization of the shelves and products, giving an invaluable tool for retail marketing teams. _x000D__x000D_As an envisaged outcome of the project, the project members – building on their existing technologies – will elaborate an integrated product that will serve the optimal placement of goods and advertisements in retail shops. Ps will also join in the marketing and sales activities to foster successful market distribution._x000D_
|
Acronym
|
CUBEA
(Reference Number: 4748)
|
|
Duration
|
01/07/2009 - 30/06/2010
|
|
Project Topic
|
CUBEA aims to create a novel customer behavior analysis system for retail shops, which can track shoppers by video surveillance cameras and correlate this information with their purchases. Analysis and prediction of customers' behavior helps in optimizing placement of goods and maximizing revenue.
|
|
Network
|
Eurostars
|
|
Call
|
Eurostars Cut-Off 2
|
Project partner
