Re-Identification with RaspberryPi and NCS2

Person re-identification is a computer vision technique that aims to identify individuals across different camera views or over time.
The technique has proven to be an effective solution for monitoring and tracking customer behavior in retail shops. By implementing this technology in a shop, we were able to gain valuable insights into customer preferences, behaviors, and traffic flow, which we used to optimize the shop layout, improve customer experience, and increase revenue.

Re-Identification model output


Include the benchmark results of running multiple model precisions. The CPU used was Intel® Core™ i3-8350K CPU @ 4.00GHz × 4 and 16 GB Ram

For Detection model

PropertiesCPUNCS with CPUNCS2 with Raspberry
Model Loading0.1802.149
Infer Time0.0210.033

For Re-Idendification model

PropertiesCPUNCS with CPUNCS2 with Raspberry
Model Loading0.2772.203
Infer Time0.0190.115


Our customers are our biggest fans

Hear from the clients we have worked with