PerClass Hyperspectral Software

Easy to use – Get results on your own images in minutes
Powerful automatic classification and regression without programming or machine learning expertise
Works with data from any spectral sensor
Batch process multiple cubes and export your results as images, data sets and Excel tables
Real-time deployment of solutions in custom applications with easy-to-integrate runtime
new! Tightly integrated with perClass Stage – a portable lab-scanning kit for line-scan and snapshot cameras

Description

perClass Mira is the easiest interface for spectral imaging with real-time deployment

Spectral cameras enable new types of applications sensing material composition. perClass Mira empowers you to interpret spectral images in minutes and create robust and accurate solutions.

Interpret spectral images in minutes

perClass Mira user interface enables any domain expert or R&D specialist to build solutions for automatic segmentation of spectral images.

perClass Mira live session – virus detection

In this live session, we define automatic segmentation of virus infection in potato plant. Data courtesy Wageningen WUR research.

Live session: Textile classification

Benefits:

Easy to use interface

Within few minutes, anyone can define classes and build image interpretation solutions using logical and simple-to-use interface.
No need for programming or specialized machine learning expertise.

picture shows how easy it is to understand perClass Mira hyper spectral camera

Automatic Machine Learning

perClass Mira is built using powerful automatic machine learning engine that selects the best statistical model for given labeled examples. It is build for real-world use cases that are non-linear, multi-class and multi-modal.

Works on your laptop

perClass Mira is built for commodity hardware. No need to purchase a powerful workstation to use it. Our intention is that any biologist or plant specialist can use it on her existing machine.

Save time with Active Learning

perClass Mira provides feedback suggesting the user places where additional labeling actually helps to improve the model. Stop wasting time with blind labeling that only increases training time and confuses the model.

Leverage your GPU to speed up analysis

Benefit from computational power present in your GPU. Speed up processing of large spectral cubes, and improve your insight by validating your solution on many large cubes.

Deploy in production

Solutions may be exported for execution using perClass Runtime. It is very easy to embed automatic classification into custom application. New images can be processed by models exported from perClass Mira. You can also execute classifiers directly on a live spectral video feed.
The screenshot shows integration of perClass Runtime into Cubert Utilities

Delivery and Returns Content description.