Skip to ContentSkip to Navigation
Research Bernoulli Institute Calendar Colloquia - Computer Science

Colloquium Computer Science, Prof. Dr. Enrico Grosso (University of Sassari)

24 February 2016

Date:                      

Wednesday, February 24th 2016

Speaker:

Prof.dr. Enrico Grosso

Room:

5161.0267 (Bernoulliborg)

Time:

15.00

Title: Face analysis: from similarity to identity

                           


Abstract:

Face identification and face verification are among the most successful applications of image analysis and understanding; this explains the significant attention face analysis has received during the past several years. This trend is largely motivated by the wide range of commercial and law enforcement applications requiring personal authentication, but face analysis is much more, as it gives access not only to the basic recognition of individuals but also to several important types of social information like categorization, kin recognition and detection of emotional cues.

This talk is mainly about face recognition, which is often related to 2D image matching but which also significantly differs from 2D image matching for a number of reasons. In fact, even though current face recognition systems have systematically improved up to a good level of maturity, their application in unconstrained environments is still challenging and this is mainly because 2D image matching techniques fail in the conditions imposed by many real applications. From this perspective, current systems are still far away from capabilities of humans.

The problem is tackled focusing first on the human perception of similarity and showing that, despite its fundamental role, measuring perceived similarity between faces is not trivial from both a theoretical and methodological point of view. Then, a special case of similarity judgment, that is kinship, is investigated in order to understand whether the perceived similarity is the only criterion adopted in judgments of kinship and similarity, or whether additional both visual and conceptual cues are used. Finally, the talk argues about computational models for face recognition and discusses some recent hybrid/multi-stage methods, combining holistic, feature-based and 3D approaches, that recently allowed to increase the efficiency and the robustness of the identification process.

Colloquium coordinators are Prof.dr. M. Aiello (e-mail : M.Aiello rug.nl ) and
Prof.dr. M. Biehl (e-mail: M.Biehl rug.nl )

http://www.rug.nl/research/jbi/news/colloquia/computerscience

Last modified:07 June 2018 10.59 a.m.

More news