IHPostal’s stamp recognition system provides an automatic and reliable analysis of postage data in mails.

Combining advanced image analysis techniques with state of the art machine learning algorithms to deal with the most difficult cases, we can identify correctly more than 98% of the stamps.

Multiple stamps per mail item can be correctly detected despite of their orientation, truncations and/or occlusions caused either by the sender of the mail item or by the sorting process. Thanks to highly optimized algorithms, our system ensures the online processing of big amounts of items.

The system is continuously refined and improved by our research department, and can be easily updated by the customer when it is necessary to add new stamps to the database. It will lead to considerable cost savings, thanks to the automatisation of the process as well as the possibility of efficiently identification of mail items with insufficient postage.

Stamps or Meter Mark are found on mail items

The Area of interest are selected

The actual postage are read as well as the Meter Mark license number