Skip to Content
Technical Articles
Author's profile photo Darius Heydarian

Get Your First Inventory Item Picture Recognition App Live in 15 Minutes

It all began on a co-innovation project together with an SAP Partner and his customer which needed to identify spare parts, with no serial number or barcode printed on it. After the prototype was build, the core functionality of the prototype has been transformed into an open source sample code to get a head start in your projects and build your own intellectual property around it.

In the spare part recognition sample code, the core functionality got divided in three steps:

  1. Synchronize all item data from SAP Business One or SAP Business ByDesign into SAP Business Technology Platform
  2. Use Tensorflow to extract the features vectors of a given input product image
  3. Take or upload a picture of an item and find out smiliar items based on a pretrained base model (and pretrained weights) provided by the ImageNet dataset for matching and retrieving related item codes (along with other data, such as quantity and price)

See the video on how to upload the Item Picture recognition sample code configured for your SAP Business One/ ByDesign backend.

Are you interested to co-innovate with SAP and having the need of an intelligent technology, such as machine learning, IoT, Blockchain, you name it? Please get in contact with your SAP Business One or SAP Business ByDesign Local Product Expert at SAP.

Screenshot of the prototype:

Assigned Tags

      1 Comment
      You must be Logged on to comment or reply to a post.
      Author's profile photo Michael Renz
      Michael Renz

      Very well done and nice to see how a ready to use code sample evolved from an initially simple PoC. Another fun fact: If I'm counting correctly this project is an outcome of the collaboration of at least 4 nations, another prove that diversity fuels innovation!