[ RobotMonkey Software Home ]

  Software Developer Peter Glen has demonstrated a simple but effective voice recognition algorithm that rivals currently existing systems both in performance and accuracy.

   The system works by decomposing elements of the speech.  The decomposition provides  basis for the recognition, and recognition is achieved by a routine called  'doughnut' , which - following the traditional wisdom - suggests to 'Look at the doughnut , not the doughnut hole'  ...  a way to analyze dominant presence of the parts of the speech.  The recognition excels  in achieving speaker independence, microphone independence and noise tolerance. In its current form it is suited for simple tasks like commanding a medical workstation to show 'medications' screen by issuing a voice command, or receiving numbers by means of voice interaction.

 One of the outstanding features of the algorithm is 'push-ability'. When the user pronounces a word more distinctly, the algorithm responds with better recognition, as opposed to the traditional algorithms, which degrade when someone is enunciating harder (or 'pushing it').  The algorithm also uses less processing power than its predecessors. This opens up voice recognition for embedded platforms, smart phones and the like. The current implementation uses 5-15% of a 2.0 GHz  AMD processor in bursts.

 The developer is seeking assistance in further developing the algorithm, and seeking potential embodiment for usage, application platforms and licensing.