POSSIBLE PROJECTS
-
1. Face detection: vision algorithm for detecting human faces from
a video sequence containing muliple people.
-
2. Directing microphone: combine face detection with "pointing"
a microphone at the current speaker. This can be done either in software,
using a microphone phase array (first building the theory on a simulation),
or using a pan/tilt unit under active control of a computer.
-
3. Modeling two-person, same-place collaboration.
-
4. Semantic representation (ontology) of a particular domain. (Campus
map, restuarant ordering, web browsing, etc.)
-
5. Speech/gesture dialog system for a particular domain. Using text/mouse
as a starting point for simulating the interaction then incorporating it
in the iMAP system.
-
6. Glove-based interface. Possible two-handed gestures for manipulating
a graphical object.
-
7. Role of human memory modeling for spatial interaction.
-
8. Analysis of speech/gesture data from a natural interaction domain.
iMAP or weather domain. relative timing, co-occurrence, grammar,
etc.
-
9. Analysis of multimodal interface in a GIS system.
-
10. Analysis of speaker-dpendency on small-vocabulary speech recognition
system.
-
11. Role of gaze in a multimodal interface.
-
12. Visual tracking of gaze.
-
13. On-line gesture recognition using HMMs.
-
14. Formulation of multimodal dialog responses of a multimodal system.
-
15. Browsing a video repository by speech/gesture natural interface.
-
17. Multimodal interface in an Augmented Reality System.