This Small Business Innovation Research (SBIR) Phase I project aims at developing video mining techniques for automatically generating customer behavior statistics to help retail enterprises. These statistics can be very valuable for supporting critical decisions in merchandising, in-store marketing, and customer service. This video mining tool can be used for accurate assessment of the effectiveness of all consumer-facing elements in retail environments designed to promote or sell products. The key problems addressed will include view-independent person detection, multi-person tracking, a method for specifying behaviors, and robust behavior recognition. The approach will be to use a variety of computer vision and statistical learning techniques under the constraints of a typical retail environment. An experimental prototype will be implemented in Phase I and will use actual surveillance videos collected from retail enterprises to establish the feasibility of the approach.
Tough economic times and hyper-competition in the retail sector demand a sense of fiscal discipline and resource optimization. One key element of improving the performance of the competing retailers and manufacturers is in developing a deeper understanding of in-store consumer behavior. Current methods (human observation and manual video indexing) for analyzing customer behavior are limited in the kind of customer knowledge generated besides being expensive and time-consuming. This SBIR Phase I effort has the potential to significantly impact the use of technology in retail business process optimization. The insights gained in customer behavior by the use of the video mining tool will enable more informed decision-making for in-store marketing; merchandise placement, and customer service. The project can impact the field of observational research in general, and the spillover benefits could go to other areas like video surveillance for loss prevention and homeland security.
