Home
Why Mobile?
Mobile Speech Recognizer
Persuasive Health Information
Mobile Language Learning

Berkeley Institute of Design

Windows Mobile Speech Recognizer

We have ported CMU's Sphinx-3.5 open-source recognizer for large-vocabulary continuous speech recognition to the Windows Mobile platform.  We are using a fixed-point MFCC front-end, and have accelerated parts of the Gaussian evaluation in assembly. This significantly increases the recognizer's performance such that it runs in real-time on Windows mobile smartphones. 

The recognizer is available for download: Download SphinxTiny

Divya Ramachandran
divya At eecs.berkeley.edu

Matthew Kam
mattkam At eecs.berkeley.edu

John Canny
jfc At eecs.berkeley.edu

Seth Horrigan
eomer At cs.berkeley.edu