Gradiant in the 12th ECCV, European Conference on Computer Vision

During the week of 7 to 13 October, the touristic city of Florence held the 12th ECCV, European Conference on Computer Vision one of the most impact conferences in the field of computer vision. More than 1200 researches around the world met themselves in the tuscan city for present their latest research works, for see the different tutorials and demos, and look for the emergent works shown in the associated workshops.


The conference was divided into three days of tutorials and workshops, and another four days for presenting the 368 accepted posters, 40 oral presentations and 22 demos. The number of posters and oral presentations accounted a 28% rate of acceptance over the whole works presented. Over all the papers, a significative percentage were related to works about scene segmentation and new features extraction, although there were works in face alignment, object recognition, scene understanding and facial expression recognition, lines of interest for Gradiant.


Among the demos, it is worth remarking the one presented by the Machine Perception Lab,, from the University of California in San Diego. The group, leaded by Marian Bartlett, presented an expression and action units recognition system, also with gender and smile recognition. Through this system, the MPLab shown an application with which expression synthesis and recognition are able to be trained. This helps, as they show, to improve this capabilities in autism patients. It is worth noting that this work received the best demo award.

This kind of works are in line with the Affective Computing line in which Gradiant is working. An example of this research line is the smilemeter developed in the center.

During the conference presentation, people could see that a significative percentage of works was from USA, whilst the second “country” with most contributions was Microsoft, with around a 7%. Spain provided less than 1% of contributions, which almost all corresponded to CVC de Barcelona..

Despite the high competence, Gradiant was represented in the ECCV, European Conference on Computer Vision. The researcher Enrique Sánchezwent to present his work “Continuous Regression for Non-Rigid Image Alignment”, “, developed during his stay in the  Robotics Institute (Carnegie Mellon University). In addition to his presentation, Enrique had the opportunity of meet some of the most prestigious researches in the Computer Vision field, as well as meet what products and research lines are getting more attention from the scientist community.