Profile Update 28.02.2019

Current activities:

I have just started the fourth year of my PhD and agreed with my advisor Prof. Peter König on a 7-months extension, that is I will aim at submitting my thesis by October 2019. Currently, I am working on analyzing recent promising results related to the main project of my PhD, about the advantages of data augmentation for deep learning and computational neuroscience. I will probably start writing at least one manuscript soon and we hope to develop the project further on and discover new interesting things that would help finish my PhD in a much better position. Importantly, I recently submitted the main paper of my PhD to the International Conference on Machine Learning (ICML) and whether it is accepted or not will have a great impact on our satisfaction.


In terms of findings, I am reasonably happy that I was able to show, both empirically and theoretically, that data augmentation has some important advantages over popular explicit regularization techniques for deep neural networks, as well as showing its potential to yield better computational models of the higher visual cortex. In terms of publications, although I have not managed yet to publish a paper in one of the top, highly competitive machine learning conferences, I did publish a paper at the International Conference on Artificial Neural Networks, the major European conference on neural networks, where I won the Best Paper Award. I also presented a short paper at the Cognitive Computational Conference, about the similarities between artificial neural networks and the inferior temporal cortex.

Future plans:

My immediate plans are to take my project as far as possible before I officially finish my PhD. After that, my plan is to stay in academia and therefore start a postdoc at a different laboratory.

My NextGenVis Experience:

Being part of NextGenVis has been a wonderful experience. In my case, I joined quite late compared to the rest, came from a totally different background and my research is on a different field, unfortunately I could not enjoy all the benefits of being part of such an amazing network of people and institutions. Nonetheless, it encouraged me to learn a lot of new things and develop an interdisciplinary profile that I am proud of and I hope will pay off in the future. At a personal level, I can only be grateful for knowing very nice, intelligent people, some of which I can now call friends.


PhD Student in Cognitive Science

  • University of Osnabrück, Germany
  • WhiteMatter Labs GmbH, Berlin, Germany


  • B.Sc. Audiovisual Systems Engineering, University Carlos III of Madrid (2009-2014).
  • M.Sc. Multimedia and Communications, University Carlos III of Madrid (2014-2015).

Previous research experience

  • Researcher and teaching assistant at the Group of Multimedia Processing in the Department of Signal Theory and Communications, University Carlos III of Madrid (2013-2016).
    • Research topics: automatic video aesthetics assessment, video affective content analysis and image processing.
    • Taught courses: Digital Image Processing, Multimedia Information Processing and Digital TV.

Research interests

  • Machine learning
  • Deep neural networks
  • Computer vision
  • Human visual affective perception

Current NextGenVis project

Predicting visual emotional perception through artificial neural networks: the main goal is to develop machine learning models based on deep neural networks that are able to predict the elicited emotional and attentional response of images on their viewers. For that purpose, I research into ways of combining semi-supervised learning models, network visualization techniques and eye-tracking data that help overcome the problem of the lack of enough labeled data.

Besides, I am involved in a project about developing a model to learn global saliency (attention) scores from images, departing from eye-tracking data, that would ultimately serve as a diagnostic tool of Autism Spectrum Disorder (ASD).


  • A. Hernández-García, F. Fernández-Martínez, F. Díaz de María (2016). “Comparing visual descriptors and automatic rating strategies for video aesthetics prediction”. Journal Signal Processing: Image Communication.
  • F. Fernández-Martínez, A. Hernández-García, F. Díaz de María (2015). “Succeeding metadata based annotation scheme and visual tips for the automatic assessment of video aesthetic quality in car commercials”. International Journal Expert Systems With Applications
  • F. Fernández-Martínez, A. Hernández-García, F. Díaz de María, A. Gallardo-Antolín (2014). “Combining audio-visual features for viewers’ perception classification of Youtube car commercials”. Workshop on Speech, Language and Audio in Multimedia (SLAM).