Visualize and interact with your data

Understanding your data enables you to best choose the classification method that fits your problem. perClass provides you with powerful interactive tools. You can now easily get a feeling from your data, leveraging all the available information, visualizing subsets of interest, painting sample labels interactively, and much more...

See videos

 perClass Interactive scatter plot
Mixture of Gaussians automatically estimating number of components.

Train generalizing classifiers

If your problems are large or small, you want classifiers that are accurate and robust. perClass combines the flexibility of Matlab design with the speed of C implementation, scalable to large industrial problems. We strive to make things simple for you, so that even complex tasks can be performed easily. Learn more

See videos

Optimize your classifiers with ROC

You know best what are the requirements of your classification task. perClass enables you to easily tune your classifiers, build robust detectors, and optimize classifier hierarchies. Our in-house research provides you with cutting edge solutions and unique tools. Learn more

See videos

 Optimizing trained classifiers with multi-class ROC anaysis
Execution of trained classifiers in applications with perClass runtime library

Quickly deploy your solutions

Do you want to run a classifier in your application? With perClass you can do it straight away! It takes you only few minutes to get your classifiers running in real-time applications. Cut your development costs and focus on what really matters: making great classifiers.

Learn how and see a short video