Login Register

Auto-login on future visits

Forgot your password?

Newsletter April 2013

perClass on Raspberry Pi!


demo running a gender classifier in raspberry Pi

We're happy to announce an experimental port of perClass Runtime to Raspberry Pi (ARM Linux, $25 credit-card size computer). It allows developers to quickly deploy any classification algorithm, trained in perClass, to Raspberry Pi devices.

This opens an entirely new set of possible applications ranging from face and person recognition (gender, age), person identification to robot vision tasks.
Are you using Raspberry Pi? Would you like to embed classifiers into your applications? Request a demo of perClass for the Pi!
Read more about the classification speed on the Pi.

A remote sensing example of clustering

Define ground-truth by clustering


How to gather a labelled training set from images? You could paint some relevant areas manually. But this is tedious and not friendly for your wrist. This tutorial shows how clustering may provide a better alternative.

Watch this 3 minute video to find out more.

Impressions from March 2013 perClass course

Group picture of the course participants

From the blog

In our March perClass course we welcomed eight participants from several countries including Turkey, Israel and Canada. It has been a strong group, with several participants having an extensive prior experience in industrial machine learning. We had nice discussions, and were happy to exchange different views and approaches for solving recognition challenges. As everyone was of a different nationality, we also had fun sharing our perspective ranging from food to approach to life :)
We are especially happy that several participants went home with novel or improved solutions for long term unsolved problems. As one participant wrote: "Thanks again for the excellent course. I gained the right knoweldge to delve futher into our machine learning problems."

Networking event 22 April

From the blog

The annual yesDelft networking event provides a great opportunity to see the many developments in yesDelft companies. This year, we showed a demo of a real-time gender classification running both on PC and Raspberry Pi board and a real-time hyperspectral French fries defect classification. The classifiers trained in Matlab may be directly executed in the sorting application. This allows quick experimentation with research ideas instantly seeing real feedback from a machine.
We've got many visitors from different sectors ranging from industry (embedded, horticulture, IT) to research and educational institutions building their own smart systems with robotics and computer vision. Read more in our blog.

Pictures from the event