Last week we had our Autumn Applied Machine Learning course in Delft. We had a great group of eight enthusiastic participants coming from as far as South Africa. Traditionally, many diverse application areas from mineral sorting, hyperspectral tissue characterization, remote sensing tree species classification, automotive, predictive maintenance, food authentication and plant biology.
2.0 release is a major update bringing tools to understand and improve models, develop solutions on object level including classification and regression and do model validation on live data stream.
perClass Mira models can be now applied to a live stream of spectral data from Specim FX cameras. This new built-in functionality allows users to quickly validate the solution accuracy and execution speed from the GUI.
perClass Mira 1.4 brings new perClass Mira Runtime DLL capable of real-time deployment of pixel classification and object detection work-flows. In this benchmark, we measure execution speed on a realistic use-case on three different systems.
We have developed an algorithm to assess brix value (sugar content) of tomatoes remotely using hyperspectral imaging based on a data set, provided by our customer Wageningen WUR.
Happy to share that perClass Mira is now available on Linux platform.
The video of our presentation at Industrial Vision days during the VISION Stuttgart trade-fair in November 2018 is now available.
We have given a live demo of perClass Mira processing hyperspectral data at the EPRISE event in Amsterdam.
The latest 1.3 release of perClass Mira brings number of usability improvements requested by customers. See a short screen-cast!
Last week we had a Practical Machine Learning training course in Eindhoven. We had a nice group of professionals working on diverse applications ranging from audio processing to defect detection using hyperspectral imaging, vibration diagnostics, mechatronic system validation and video analysis.
This week, we presented live demo of plastic contamination detector at Agri&Food Tech event in Den Bosch, The Netherlands. In the demo, perClass Mira was used to design and execute live detection of foreign objects in food product, invisible to naked human eye, using Specim FX17 camera.
Last week, we have participated in the world's largest machine vision show in Stuttgart, Germany. perClass got a significant attention from industrial practitioners, system integrators, distributors and camera manufacturers.
Last week, we gave the 19th edition of Machine Learning course for practitioners. We have discussed a broad range of applications including defect detection, geophysical applications, medical support systems and remote sensing.
What a busy week! Another event related to food processing, this time in Brightlands Venlo. Live demo of spectral image interpretation gave us a unique opportunity to understand challenges of practitioners from food production in this part of The Netherlands.
We gave a live demo at specialized event on Photonics for Agro&Food indutry.
perClass sponsored SpectroExpo event, part of WHISPERS conference on hyperspectral imaging in Amsterdam. It was a great opportunity to meet both our industrial partners and users of spectral imaging systems.
Happy to announce that SpectraPartners became a distributor of perClass software in Benelux!
Users of Imec spectral cameras may now benefit from a tight integration between perClass Mira and HSIViewer.
New release of perClass Mira brings many enhancements such as confusion matrix visualization of error structure, interactive performance fine-tuning, and RGB preview mode.
The research team of RECENDT has published prrogress on their new mid-infrared hypespectral scanner illustrating material separation using perClass Mira
Happy to share a publication made by Wageningen WUR researchers who used perClass Mira together with Specim I camera to validate authenticity of bananas.
Together with SpectraPartners, distributor of Specim spectral cameras, we have given a joint demo at Wageningen WUR Summerschool of plant phenotyping. Participants could see spectral image acquisition and interpretation in minutes without programming.
The video of our presentation at Chii 2018 hyperspectral workshop in Graz is now available.
We are happy to announce our Machine Learning autumn course. You are welcome to bring your own data for the practicum. The course includes hands-on exercises on Deep Learning. Would you like to join? Register now.
Creating classification solutions for hyperspectral images is now only a matter of few minutes. In this short video Cubert illustrates the workflow for spices classification. In perClass Mira the user can easily create a classification solution. Cubert has integrated perClass Runtime in their camera acquisition software. In this way the classification model can be applied directly to the live camera feed.
We have participated in the Chii 2018 hyperspectral imaging workshop in Graz, Austria. Happy to share huge amount of interest in our perClass Mira user interface.
perClass was represented at the Vision & Robotics trade-fair in Veldhoven, The Netherlands at the stand of SpectraPartners.
We have welcome a diverse group of researchers working on applications in industrial quality control, agro & food robotics, traffic and maintenance, brain computer interfaces and multimedia content analysis.
At Agri&Food fair, we have demonstrated the first public beta of the perClass Mira environment for user-friendly interpretation of spectral images.
We’re happy to announce a partnership with Cubert GmbH a pioneer in real-time spectral imaging. Via this partnership, perClass comes tightly integrated into Cubert software and directly useful for all Cubert customers. I comes as a part of standard Cubert Utilities installer.
In the last week course, we had welcomed participants from Netherlands, Belgium, UK and Spain.
University Liège has been using perClass to build a metal sorting system based on hyperspectral imaging. Classifying metals in hyperspectral images is very challenging task
Last week, we participated in the Vision Trade Fair in Stuttgart, Germany, the world’s leading machine vision event. It was our fourth exhibition and, by far, the most exciting. While few years back, we had to explain what we mean by Machine Learning, today’s environment and public perception entirely changed.
In our October course, we have welcomed a group of 10 researchers from The Netherlands, Mexico and Germany. Participants brought in a lot of enthusiasm and interest especially in Deep Learning.