Computer Vision Solution for In-Store Customer and Employee Behavior Analysis

About Our Client

The Client is a recognized authority in the production of collaborative category management and store mapping solutions.

Challenge

The Client wanted to develop a solution enabling a retailer or distributor to analyze consumers’ behavioral patterns, more effectively plan merchandizing activities, estimate work of sales personnel, thus resulting in sales and profit increases.

Solution

A custom software development company with 32 years in IT and 8 years in image analysis, ScienceSoft created an innovative computer vision solution for the retail industry consisting of the following components:

• A set of camera software (image capturing, camera setup and calibration, camera status monitoring);
• People tracking module (tracking of store staff and customers, using fish-eye cameras);
• Shelf tracking module (tracking of hand movements);
• Fridge tracking module (tracking of hand movement, when interacting with glass-door fridge);
• SilverLight portal (access point for all gathered data and analytics).

Results

ScienceSoft’s team successfully solved the challenging task and promptly delivered the custom image analysis solution that can gather a large amount of merchandising information to perform:

• Analysis of store’s staff activities and intercommunication with customers;
• Tracking of the customers’ movements and activities, including counting customers in specific areas;
• Queue analysis (average number of customers in line, average waiting time, leave count, etc.);
• Hand movement tracking (detection of most popular shelves, products, places);
• Merchandising data collection and analysis.

Technologies and Tools

Intel Parallel Studio, Intel C++ Compiler Professional, OpenCV, Integrated Performance Primitives, NVIDIA CUDA.