I
Irene Kotsia
Researcher at Middlesex University
Publications - 61
Citations - 5327
Irene Kotsia is an academic researcher from Middlesex University. The author has contributed to research in topics: Facial expression & Facial recognition system. The author has an hindex of 24, co-authored 61 publications receiving 3405 citations. Previous affiliations of Irene Kotsia include Aristotle University of Thessaloniki & University of London.
Papers
More filters
Proceedings Article
Higher order support tensor regression for head pose estimation
TL;DR: The model is based on the Canonical (CAN-DECOMP)/Parallel Factors (PARAFAC) decomposition of tensors of multiple modes and allows the simultaneous projection of an input tensor to more than one discriminative directions along each mode.
Book ChapterDOI
Exploring the similarities of neighboring spatiotemporal points for action pair matching
Irene Kotsia,Ioannis Patras +1 more
TL;DR: A novel similarity measure is presented that exploits the underlying relationships among neighborhoods of detected spatiotemporal points in a frame of an image sequence to create a similarity vector that can be used as an input to a classifier.
Recognition of facial expressions in presence of partial occlusion
TL;DR: This work is interested in finding which part of the face comprised sufficient information with respect to the entire face, in order to correctly classify these six basic facial expressions when the eyes and eyebrows or the mouth regions are left out.
Book ChapterDOI
Facial expression recognition using shape and texture information
Irene Kotsia,Ioannis Pitas +1 more
TL;DR: The Discriminant Non-negative Matrix Factorization (DNMF) algorithm is applied at the image cor-responding to the greatest intensity of the facial expression (last frame of the video sequence), extracting that way the texture information.
Proceedings ArticleDOI
Support tensor action spotting
Irene Kotsia,Ioannis Patras +1 more
TL;DR: This paper first calculates a novel objective function between an input video sequence and an action's weights tensor, as acquired from a Support Tensor Machine classifier, and searches for an appropriate transformation that maximizes the objective function.