University of Birmingham > Talks@bham > Artificial Intelligence and Natural Computation seminars > Compositional Object and Activity Recognition

Compositional Object and Activity Recognition

Add to your list(s) Download to your calendar using vCal

If you have a question about this talk, please contact Lars Kunze.

Host: Ales Leonardis, Krzysztof Walas

Abstract: In recent years there has been an increasing interest in compositionality and mid-level representations, since they effectively bridge the large semantic gap between pixels and object categories. The main paradigm in this area is an orderly partonomy based on few atomic parts. We pursue the opposite strategy by randomly sampling a large number of instance specific part classifiers. Due to their number, directly combining all parts in a single classifier is not feasible. Therefore, we follow a max-margin approach and seek randomized compositions that are discriminative and generalize over all instances of a category. The approach not only localizes objects in cluttered scenes, but also explains them by parsing with compositions and their constituent parts. Similarly, latent compositions have proven successful in video analysis, particularly when modeling collective activities of a group of persons. To model interactions and context, meaningful activity constituents are obtained not merely based on visual similarity but based on the function they fulfill on a set of validation images. Time permitting I can also discuss recent applications of this work in neuroscience and the digital humanities.

Speaker’s homepage:

This talk is part of the Artificial Intelligence and Natural Computation seminars series.

Tell a friend about this talk:

This talk is included in these lists:

Note that ex-directory lists are not shown.


Talks@bham, University of Birmingham. Contact Us | Help and Documentation | Privacy and Publicity.
talks@bham is based on from the University of Cambridge.