The CiTIUS celebrates once again the recognition of another thesis from the center in the Extraordinary Doctorate Awards of the University of Santiago de Compostela. Following the granting of this award to two doctoral works in the previous edition, the thesis of the CiTIUS graduate researcher Lorenzo Vaquero adds to this trajectory of excellence and also receives support for his doctoral work, developed at the CiTIUS of the University of Santiago de Compostela. The award, of an annual nature, is aimed at those doctoral works rated with the highest score by the tribunal, which have also received the Cum Laude mention and a score of 12 points or higher in the evaluations issued by its members. Under the title Visual Multi-Object Tracking through Deep Learning, and under the supervision of CiTIUS researchers Manuel Mucientes Molina and Víctor Manuel Brea Sánchez, the thesis presents deep learning solutions for real-time tracking of multiple objects in videos, an essential component in applications such as robotics or video surveillance.

The work proposes innovative approaches to improve traditional detection systems, which depend on frame-by-frame systems and geometric attributes, which are not suitable for real-time environments, nor in the case of open-set scenarios. To address the limitations of these conventional systems, Lorenzo Vaquero's doctoral work led to SiamMT, an architecture that adapts single-object tracking techniques to handle multiple targets. Subsequently, this idea was refined with the introduction of SiamMOTION, which improves performance by managing actors and accommodating objects of diverse sizes. Finally, an architecture based on Transformers was introduced that complements a lightweight detector, enhancing the retrieval of undetected objects and the overall tracking performance.

It all started with a summer scholarship...

The beginning of Lorenzo Vaquero's research career dates back to his participation in the CiTIUS Summer Scholarships program, which allowed him to delve into the research field that ultimately led to his thesis. Previously, during his undergraduate and master's training, Lorenzo specialized in the fields of Computer Vision and Artificial Intelligence, receiving two additional awards during this formative stage: the national award for mobility and accessibility from the Renault Foundation and the 'Club de Excelencia en Sostenibilidad' (Best STEM Final Degree Project) and the DXC Spain award for the best Master's Final Project (TFM) from the Interuniversity Master's in Big Data. Currently, Lorenzo remains connected to the CiTIUS as an external collaborator researcher and works as a postdoctoral researcher at the Fondazione Bruno Kessler in Trento, Italy, where he manages various international projects focused on human-centered computer vision research.

The researcher emphasized "the boost that this award represents to continue exploring new frontiers in the world of technology and science," expressing his satisfaction with this recognition. "I feel immense excitement and gratitude," he added, "for continuing to collaborate with the wonderful CiTIUS community." Reflecting words filled with praise also included expressions of gratitude to his supervisors: "I want to express my deep appreciation to USC, to CiTIUS, to my thesis directors, Manuel Mucientes and Víctor Brea, as well as to all the colleagues and friends who made this journey possible." The award will be presented by the International Doctoral School of USC (EDIUS) during a ceremony scheduled for next June.