Start-Up of the Month: Fizyr

What was the idea that initiated the foundation of your company?

It started as Delft Robotics, a spin off from Delft University by Prof. Martijn Wisse.

(Bild: Fizyr B.V.)

The aim was to integrate robots for the SME. In 2016 we participated in the Amazon Picking Challenge and by applying deep learning we won both the picking and stowing competition. We realized that we like to apply the latest academic AI technologies to solve big problems of the industry. That led to our pivot in 2017 of only computer vision and machine vision software by using deep learning.

What is the goal you want to reach?

In our DNA we are very ambitious and want to become and/or stay #1 in computer vision for certain applications. We currently do most of our work in logistics meaning enabling robots to cope with huge variation in shape, size, colour, material etc. We have created solutions for item/order picking in e-fulfilment as well as for picking unknown parcels for the courier industry and we want to keep serving the global industry with the best possible AI algorithms.

What is the question your product/technology is the answer for?

Computer and machine vision systems can have a hard time dealing with variation or for instance find and classify unknown defects. We train algorithms to take over repetitive tasks from humans, often faster, more accurate and very predictive. The key first step is object segmentation, secondly is classification which enables counting.

What makes your company unique?

We have a brilliant team that combines the necessary technical and business qualifications and experiences that enables us to provide the required long-term solution for corporates. We have a lot of fun with the team and frequently do activities together. Since we have colleagues from six different countries everyone speaks in their second language English.

Video: One minute summary of what Fizyr solves in logistics