research interests overview
We are looking for ph.d. students in three projects:
- a. Linguistic annotation and search
in images
and video, funded by Samsung R&D
- This project envisages generating linguistic descriptions and
annotations for large image / video datasets. The project will require the
student to work in the area of machine
learning, and particularly deep
learning methods such as CNN and Long-Short-Term Memory (LSTM)
networks.
Jointly with Prof. Vinay Namboodiri.
- b. Spatial Navigation and Cognitive Map learning - project
with University of Melbourne.
- Navigation in indoor and outdoor spaces,
building cognitive map like structures from sensory data, generating
descriptions to guide others. Will involve
machine learning techniques for dimensionality reduction, Vision and
Language integration for route description generation, etc. In
collaboration
with Prof. Stephan Winter
- c. Visual vehicle driving, funded
by Center
for AI and Robotics
- Detecting anomalous events from an autonomously driven car, and also
construct a summary of the scenes.
Will involve action modeling via a probabilistic topic model and
also action recognition using CNN / RNN. see
[this WACV-15 paper]
Joint work with Prof. K.S. Venkaesh
Update: Oct 2009