Problem Statement

Learn how to track inventory, and make the entire inventory tracking process in a large warehouse autonomous.

Key Approach

We have analysed the SKUs of a large warehouse to find out unique items at any given point of time.


Extract video into frames


Preprocess each frame to enhance object detection

Build Model

Using HOG and SVM to build item recognition model


Provide reports on unique items

Business Benefits

  • Item count and tracking solutions will help to monitor inventory status and movement.
  • It will help to prevent any pilferage and alert inventory managers on a realtime basis.
  • Our solution will not intrude the privacy of the employees
  • Our algorithm can count the no of unique items with an accuracy of more than 92%.