I started my Engineering degree in 2012, in the Universidad de Zaragoza. Then, in 2017, the 2-year Masters on Engineering followed. In my second year, I participated in the Erasmus+ program and went to the RWTH Aachen University (Germany) were I started to learn computer vision and deep learning. The Erasmus was followed by an internship at Universidad de Zaragoza, in the Department of Computer Science and Systems Enginnering.

In December 2019, I presented my Master's Thesis on omnidirectional images under the supervision of my advisors (Jesús Bermudez-Cameo and Josechu Guerrero), which was the results of the internship in the department.

My Ph.D started in January 2020 in the same department and with the same advisors with which I did my Master's thesis (Why should I change something that works?). At that moment, my life as Ph.D student started. Though, the starting time was not the best 🦠😷🦠😷🦠, it allowed my to focus on publishing with less distractions (also called life). In December 2023 I finally read my thesis Learning the surroundings: 3D scene understanding from omnidirectional images, achieving the calification of cum laude.

My research interest is on Computer Vision and Deep learning, and more specifically, scene understanding from omnidirectional images. I am also interested on geometrical and photometric approaches inside the computer vision field. I do have some experience on geometrical solutions (with great help of my advisors), but I have mostly focused on deep learning solutions in these last years. In my research stay in the CNRS-AIST JRL (Japan), I also got many help from fellow researchers on photometric approaches and with the combinations of methods I started in the field of practical applications of omnidirectional cameras.

Here I present a collection of the publications where I participated and the available code of some of them.

• 2023:

  • Distortion-aware neural networks:
    • Conference article, BMVC'23: Convolution kernel adaptation to calibrated fisheye
      • Code: Calibrated Convolutions
  • Semantic segmentation and depth estimation:
    • Conference article, ICRA'23: FreDSNet: Joint Monocular Depth and Semantic Segmentation with Fast Fourier Convolutions
      • Code: FreDSNet
  • Omnidirectional image stabilization:
    • Workshop article, CVPR'23 workshop OmniCV: Visual Gyroscope: Combination of Deep Learning Features and Direct Alignment for Panoramic Stabilization
      • Code is in different repos. First part of the pipeline is: HoLiNet

• 2022:

  • Layout estimation from non-central panoramas:
    • Journal article, Pattern Recognition Elsevier: Atlanta Scaled layouts from non-central panoramas
      • Code: Scaled layouts
    • Dataset article, Data in brief Elsevier: Non-central panorama indoor dataset
      • Code (from Jesus Bermudez-Cameo repository): Non Central Indoor dataset

• 2021:

  • Layout estimation from non-central panoramas:
    • Workshop article, CVPR'21 workshop OmniCV: Scaled 360 Layouts: Revisiting Non-Central Panoramas
      • Video presentation: Video

• 2020:

  • Omnidirectional image generator from Unreal Engine 4 virtual environments:
    • Journal article, Sensors MDPI: OmniSCV: An omnidirectional synthetic image generator for computer vision
      • Code: OmniSCV
      • Dataset: Omnidirectional images dataset
    • Conference article, Jornadas I3A 2020: Omnidirectional Image Data-Set for Computer Vision Applications
      • Conference poster: Poster.pdf
  • Floor detection and extension for autonomous guidance:
    • Conference article, ICARCV'20: Floor extraction and door detenction for visually impaired guidance

If you have come so far as to be reading this, that means that you already know about my GitHub profile .

You can also find my research work on Google Scholar and Orcid

If you want to contact me directly, my email address is: [email protected]

I do hope this page helps someone to get to my work (and of course, that my work is useful for many many people). And also, to add some "color", here is one of the results of my research with non-central panoramas. Enjoy it hehe.