- Less, expensive, manual handling necessary
- We can offer a read rate improvement of 30%
- Even the labour intensive Asian parcel stream can be turned into a profitable revenue stream
Many mail pieces can’t be processed automatically because the addresses are not complete or correct, damaged or not visible. This results in expensive manual handling. Our OCR is a reliable solution for automatically reading address labels. This translates to higher performance for our clients.
In the Netherlands only 6.2 million mail pieces are processed each day. Many pieces can’t be processed automatically because the addresses are not complete or correct, damaged or not visible. Those items now have to be handled manually. This is much more expensive while posts are looking for ways to reduce costs as a result of decreasing volumes.
Automation is a key factor in cost reduction. Prime Vision offers several ways to increase the read rates and decrease the error rates on letters and parcels. For handwritten items we developed the Robust Handwriting recognition engine. This engine is deep learning based and can improve the read rate with 30%. The advanced Machine Print Address Recognition engine, reduces the amount of pieces that have to go to video coding with 10%. When the postal code and city are missing, the street and addressee are read.
Challenging mail nowadays is the stream that comes from China. The addresses are, most of the time, hard to read because of bad printing of the labels and the different size and shape of the parcels. Using the latest AI techniques, we manage to read this difficult mail.
Our Optical Character Recognition (OCR) Engine is a comprehensive solution; complete with recognition process chain. It incorporates region-of-interest finding capabilities as well as the ability to interpret images and addresses.
The Recognition Engine is designed based upon Prime Vision’s core technology in the field of character recognition and machine-vision. As part of an ongoing research program, we invest heavily in the continuous development of Vision Technology, ensuring that we maintain our position as a leading global recognition developer.
By aiming to be at the forefront of recognition technology, this translates to higher performance for our clients and the capability to deliver increasingly tailored solutions.
Technology is advancing at such a rate that we can’t predict where we will be in a decade. For now, our industry’s vision is based on artificial intelligence. While the speculation of what’s possible is endless, developers have to translate all this potential, invention and innovation into realistic solutions. While it’s great to let the imagination fly, we have to balance indulgence against purpose and benefit.
Michiel de Rijcke
has led the research and development department of Prime Vision since 2001. With a solid background in computer vision and a PhD and cum laude graduation in mathematics, Michiel’s ingenuity brings applied research to innovation.