Michiel Bacchiani
The ASSM toolkit for Polynomial
Segment Models and Automatic
Unit Design
Abstract:This technical report describes the software that was produced to design and use a speech
recognition system, based on Polynomial Segment Models (PSMs). The toolkit offers a variety of
tools for the training, classification and recognition using PSMs. It also provides tools specifically
aimed at design of automatically derived acoustic units. The software described in this report
covers the available software package (ASSM) at this time, June 1997.