For vision-based gesture recognition (without dialog context), we trained a multi- class SVM with two different classes: head ... We resam- ple the velocity vector to have 32 samples per second. ... Component Object Model (COM) technology, displays the Portable Document File (PDF) and waits for the user input. ... The dialog boxes appearing during both tasks were designed to replicate reminders sent by a calendar application (i.e. Microsoft Outlook), alerts sent by an email client, andanbsp;...
|Title||:||IUI ... Conference Proceedings|
|Author||:||C. L. Sidner|