Multimodal Human-Computer Interaction
( for more information see my publications )
|
Experimental system incorporating speech, gaze, and lip-movement modalities. Automatic speech recognition provides natural conversational command selection. Lip reading is able to be utilized for enhancing the accuracy of interpretation of speech input through temporal analysis of lip images. Gaze tracking delimits a screen region or indicates one menu item by looking at the object(s) of interest in the display. |
|
Four key components are incorporating in parallel in order to interpret the subject's intention automatically. This is implemented by multi-thread programming. |
EyeTalk project seeks to build intelligent multimodal user interface into a car environment so that useful human interaction with the car navigation system(CNS) can occur in a natural manner. I am the project coordinator. My responsibility is to manage development and to build eye-gaze tracking module. EyeTalk includes followings:
[eye-gaze tracking] [lip-reading] [speech recognition]
PowerPoint Slides of EyeTalk project
กก
