Multimedia Analysis

Multimedia Analysis requires complex models to extract and combine information from images, video, speech and text. We are focusing on the adaptation of Machine Learning techniques to build these models, apply them in applicative settings, and evaluate their performance. In particular, we are now studying the use of Deep Neural Networks for creating and combining image and text representations. EURECOM is actively involved in international evaluation campaigns (e.g., TRECVID organized by the US National Institute of Standards).

Here is a list of research topics pertaining to these activities:

  • Deep learning

  • Image retrieval

  • Semantic interpretations of images


Syndicate content

Data Science