My research interest include:

1. Semantic Models: How do we represent link between various concepts ? or How do we represent and store relationships ? I am interested in exploring semantic models for visual inference. These semantic models store relationships between different concepts (nouns  and verbs). I have explored storyline models which represents possible storylines for given video(CVPR 2009). Using this model, instead of inferring only concepts we simultaneously infer concepts and causal links between the concepts. I have also looked in contextual models which are based on language itself (ECCV2008). It has been argued that language is  an instrument of thinking and therefore can represent all the relationships required for visual reasoning. In this work, we represented relationships between nouns using prepositions and comparative adjectives.
   

1.   Exploring synergy between language and vision for learning semantics: What role does language play in helping us learn semantics ? I am interested in exploring how declarative information and other linguistic information can be harnessed to efficiently learn the semantic models. I am also interesting in exploring how we can obtain such linguistic information.
   

1. Computer Vision: Activity Analysis, Object Recognition, Motion and Tracking, Camera Networks, Combining different visual cues for effective recognition.