An intelligent image-based customer analysis service

Abstract

The customer analysis can be applied widely in many situations, such as surveillance system, digital signage, store entrance, etc. In digital signage application, when customers are viewing ads, the system can automatically analyze multiple attributes of customers, including customer's gender, age, height, viewing distance, the degree of interest and total number of visitors. The advertising companies or store owners can judge potential customers who might interest in some particular products and increase the probability of the products being purchased through those customer analysis data. Another application in the store entrance is to calculate the total number of customers, gender and age statistical results in each time interval. That significant information can be used not only to integrate with point of sale (POS) system for letting the customers has a relationship with their purchased products but also to make the staff arrangement more adequate. The important equipment used in this system is the RGB-D sensor (Xtion Pro Live) which can capture the 3D depth information and 2D color image simultaneously. According to the 3D depth information, the system can detect and track the human to get the viewing distance, height and total number of visitors. Then the 2D image recognition technology is adopted to detect the face from human contour region and to recognize their gender, age and attention time for degree of interest. Finally, the system returns the analysis results and statistical information to the advertising companies or store owners for providing a reference in their business strategy.