Prev
Date:
Do you develop intelligent products that make automatic decisions?
Do you want to easily build high-performance classifiers or detectors?
Do you need to run them fast in your applications?
PRSD Studio empowers industrial and academic researchers to quickly design robust statistical classifiers in Matlab and rapidly deploy their solutions in custom applications using a DLL library.
PRSD Studio is used by researchers designing industrial sorting solutions, traffic monitoring and medical diagnostic applications.
The new 2.1 release of PRSD Studio offers tools for:
- interactive visualization of data and meta-data (e.g. class labels but also patient/object labels)
- hand-painting of labels in scatter plots and images
- cluster analysis (mixture models automatically estimating number of clusters)
- training of robust and scalable classifiers (parametric, non-parametric, neural, support vector machines, decision trees)
- construction of hierarchical classifiers and classifier fusion
- optimization of trained classifiers with two- and multi-class ROC analysis
- two- and multi-class rejection of outliers or low-confidence samples
- performance evaluation, ROC variance estimation
- cross-validation of custom algorithms (including in-training ROC analysis)
- immediate deployment of trained classifiers out of Matlab
- fast multi-threaded execution in on-line and real-time applications
- remote classifier updates in production deployments without recompilation
PR Sys Design, the creators of PRSD Studio, provide commercial support and training for industrial pattern recognition system design with PRSD Studio.
For more information see, http://prsdstudio.com