AI use cases & demos

CV-Powered In-store Customer Behavior Tracking

by Max Frolov
CEO @ DataRoot Labs
Tech Stack:
GStreamer
OpenCV
Python
YoloV2
MeDarknettal
Industry:
Consumer Behavior
Retail
Project Length:
2 months
Read independent verified review on Clutch.co


Summary

  • Insights to the customer behavior on store premises unlock enormous value to retailers.
  • The store chain client wanted to track the number of people walking in and out of her stores as well as their behavior inside the store.
  • Our team has built a solution that through a video camera tracks where people walk inside the store, identifying their gender and age category. The above had to be calculated in real-time using the setup located in the store – Dual Core Celeron + GTX 660 2Gb.

Tech Challenge

  • Since in 2016 there was no mobilenet yet, our team had to find a real-time solution for Detekces with significant capacity constraints as well as a solution for real-time video streaming with minimal loads.
  • Additionally, all data had to be cast into a dashboard displaying statistics.

Solution

  • Our solution has optimized Darknet YoloV2 and achieved 15 FPS, which was enough to solve the problem.
  • For real-time streaming we used GStreamer which we optimized for the in-store setup mentioned above.

Impact

  • As a result of our development, all meta information from the cameras was processed locally and had flown into the cloud.
  • From there, the store store chain could see the detailed statistics on the number of customers inside the store, as well as their distribution by age category and gender, allowing for smarter marketing and customer service decisions.

Updated Aug 13, 2019 — 18:00 UTC