University of Cambridge > > women@CL Speaker Lunch Series > women@CL Talklets (Session 1) - Rainbow group

women@CL Talklets (Session 1) - Rainbow group

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

If you have a question about this talk, please contact Helen Yannakoudakis.

This talklets session will feature two speakers from the Rainbow group:

Speaker: Marwa Mahmoud

Title : Automatic analysis of naturalistic hand-over-face gestures

Abstract: One of the main factors that limit the accuracy of facial analysis systems is hand occlusion. As the face becomes occluded, facial features are either lost, corrupted or erroneously detected. Hand-over-face occlusions are considered not only very common but also very challenging to handle. However, there is empirical evidence that some of these hand-over-face gestures serve as cues for recognition of cognitive mental states. In this talk, I will present an analysis of automatic detection and classification of hand-over-face gestures. The proposed approach detects hand-over-face occlusions and classifies hand-over-face gesture descriptors in videos of natural expressions using multi-modal fusion of different state-of-the-art spatial and spatio-temporal features. The detailed quantitative analysis sheds some light on the challenges of automatic classification of hand-over-face gestures in natural expressions.

Speaker: Flora P. Tasse

Title: 3D Points of Interest and Shape retrieval

Abstract: Shape saliency measures importance of points or regions on a 3D surface. It is a useful tool for several shape analysis tasks such as shape similarity, simplification, and viewpoint selection. We propose a cluster-based approach to point set saliency detection, a challenge since point sets lack topological information. Our approach detects fine-scale salient features and uninteresting regions consistently have lower saliency values. We also present shape retrieval based on salient features detected using this method. We compare this shape retrieval system to other systems that use ground-truth points and random keypoints. Results show that on average, selecting many random points on 3D surfaces produces significantly better retrieval performance than only using points of interest. In shape retrieval, salient points and non-salient points both play an important role.

This talk is part of the women@CL Speaker Lunch Series series.

Tell a friend about this talk:

This talk is included in these lists:

Note that ex-directory lists are not shown.


© 2006-2019, University of Cambridge. Contact Us | Help and Documentation | Privacy and Publicity