I am a Research Scientist at the Samsung AI R&D Center in Cambridge. From September 2016 to March, 2019, I was with the Computer Vision Lab, at the University of Nottingham. I obtained my PhD degree in Computer Science from the University of Nottingham in 2017, with a thesis on Continuous Regression for Face Tracking. In my previous life, I obtained my MSc in Signal Theory and Communications and my MEng in Telecommunications Engineering from the University of Vigo (Spain), in 2011 and 2009, respectively (GPA 8.0/10.0). I visited the Human Sensing Lab @ CMU for 6 months in 2011. My PhD was partially funded by the Vice-Chancellor’s Scholarship for Research Excellence, as well as by the School of Computer Science scholarship. Prior to this, I received a scholarship from the Fundación Pedro Barrié, to do a research stay @ the CVLab in Nottingham, in 2013.

My research interests are on the use of Deep Learning techniques for Facial Expression Recognition, as well as the fundamental theory behind the recent developments in generative methods using Deep Learning (VAEs and GANs).

GANnotation: A landmark guided face to face synthesis network (code here!)

GANnotation example

Check the pre-print paper here, where we propose a new loss to bridge the gap between the input and target distributions in GAN-based face synthesis

Heatmaps for AU intensity estimation (code here!)

Alt Text

Alt Text


(September 2019) - One paper accepted to NeurIPS 2019.

(March 2019) - In March, 2019, I will join the Samsung AI Center in Cambridge, and thus I will be leaving my position at the University of Nottingham.

(November 2018) - A new pre-print is available here : This paper presents a new loss to overcome the limitations of a self-consistency loss in GAN-based face synthesis

(August 2018) - Code for the BMVC 2018 is now available here

(July 2018) - One paper accepted at BMVC 2018. PyTorch code coming soon

(June 2018) - Our iCCR code has been developed and integrated into the iOS app Meo3d for virtual avatars

(June 2018) - Matlab code for iCCR has been updated to the GitHub account. You can find it here.

(August 2017) - Our paper “A Functional Regression approach to Facial Landmark Tracking” has been accepted at IEEE TPAMI

(May 2017) - Viva voice successfully defended

(February 2017) - Our C++ version of iCCR has been integrated into the ARIA-VALUSPA platform

(October 2016) - Our work on incremental face tracking using Continuous Regression (iCCR) has been presented at ECCV 2016

(July 2016) - Our paper “Cascaded Continuous Regression for Real-time Incremental Face Tracking” has been accepted to ECCV 2016. Check the Continuous Regression site.