Publication

pySPACE - What’s New?

Mario Michael Krell, Andrei Cristian Ignat

Bericht DFKI GmbH DFKI Documents (D) 14-07 11/2014.

Abstract

pySPACE is a signal processing and classification environment written in Python. It helps data scientist to analyze their data by providing an interface to numerous processing nodes which can be combined in a modular fashion. Furthermore, it largely automizes the parallel processing and largely simplifies the definition of processing chains. pySPACE has been developed at the University of Bremen and the DFKI, Robotics Innovation Center since 2008and is strongly connected to the in house framework reSPACE which enables the transfer of processing chains to mobile devices. In 2013 pySPACE was published and became open source. It is continuously improved. pySPACE has been presented at several conferences. The backtransformation has been added, which is a new concept to visualize and compare signal processing chains. Several tutorials have been added and a generic testing framework has been implemented. This presentation shortly summarizes recent developments in pySPACE like public relations, backtransfor- mation, regression, usage on Windows, documentation, and testing.

German Research Center for Artificial Intelligence
Deutsches Forschungszentrum für Künstliche Intelligenz