Gradiant at the Carnegie Mellon University

As part of his PhD, Enrique Sánchez, researcher at Gradiant’s Multimodal Information Area, has completed his half-time as a visitor scholar in the Human Sensing Lab at the Robotics Institute (Carnegie Mellon University, CMU), one of the most important robotics institutes in the world.

At CMU, and during six months, Enrique continues his work on Active Appearance Models (AAMs), proposing novel contributions and improvements focused on the understanding of AAMs at a fundamental level. AAMs are an important tool in face analysis, useful for facial expression analysis, face recognition or face tracking, applications which are currently being developed at Gradiant.

The Human Sensing Lab is headed by Professor Fernando de la Torre (http://cs.cmu.edu/~ftorre/). Fernando is also the leader of the Component Analysis Lab, in the Robotics Institute, and has a long experience in Computer Vision, with more than one hundred high-level international publications, and a large participation in R&D projects.

Besides his work with Fernando and the Human Sensing Lab research team, Enrique has had the opportunity to attend several talks of renowned experts in Computer Vision. One of the most interesting ones was given by Michel Valstar, from the Intelligent Behavior Understanding Group (iBug) of the Imperial College, London. The topic of the talk was facial expression recognition, with special focus on Action Unit detection, and next generation face expression analyzers. During this talk, Michel gave an overview of the most common problems reported until today in facial expression recognition, e.g., face registration or detecting emotion in non-frontal faces, and provided an overview of the results obtained in the FERA 2011 Challenge, a competition made this year on facial expression recognition (http://sspnet.eu/fera2011). These results have much interest since face expression analysis is an important research line in Gradiant, with several R&D projects focused on the development of recognition engines able to work in daily scenarios.

With this initiative, Gradiant opens the possibility to further collaborations with the Robotics Institute, another step towards the internationalization of our research center.