Abstract:
CCTV cameras are installed in almost all retail business these days. This is usually done primarily for security purposes. But in addition to security, the surveillance cameras in- stalled in retail stores could provide extra customer information that would be very helpful in improving business. This valuable information would allow marketing analysis to better understand customer behavior and provide more satisfying service. Customer experience starts with how well a store is organized, how good its ambience and facilities are, and how well a product elicits reactions of customers at appropriate times. It may also depend on the behavior of the staff. Business would thus benefit from constant monitoring and analysis of what we see in the store’s CCTV video streams. Human eyes can interpret video at a glance, but this process needs to be automated if the retailer is to obtain rich real-time insights about the store’s business. In this research study, I review on existing techniques and technologies used in video analytics for customer analysis. I also analyze the gaps in the state of the art and suggest possible improvements to existing ideas. Based on these gaps, I build and evaluate a system that monitors group activities and provides business promotional strategies using video analytics.