Pablo García Fernández

PhD Candidate

I am Pablo, a PhD candidate in Computer Vision at the University of Santiago de Compostela (USC) and a researcher at the Research Center on Intelligent Technologies (CiTIUS). My work explores the frontiers of visual recognition under minimal supervision, with a focus on object detection beyond closed-set assumptions. This includes open-vocabulary models, few-shot generalization, and category-free inference. Broadly, I aim to develop scalable, adaptable perception systems that bring computer vision closer to the richness and flexibility of human perception.

In 2024–2025, I was a visiting PhD student at the University of Trento (Italy) with the Multimedia and Human Understanding Group (MHUG).

I am funded by the FPU Predoctoral Fellowship, one of the most competitive research grants awarded by the Spanish Ministry of Universities.

My work has been published in leading venues such as ICCV.

selected publications

Superpowering Open-Vocabulary Object Detectors for X-ray Vision

Pablo Garcia-Fernandez, Lorenzo Vaquero, Mingxuan Liu, and 5 more authors

In IEEE International Conference on Computer Vision (ICCV), 2025

Abs Bib PDF Website

Open-vocabulary object detection (OvOD) is set to revolutionize security screening by enabling systems to recog nize any item in X-ray scans. However, developing effective OvOD models for X-ray imaging presents unique challenges due to data scarcity and the modality gap that prevents direct adoption of RGB-based solutions. To overcome these limitations, we propose RAXO, a training-free framework that repurposes off-the-shelf RGB OvOD detectors for robust X-ray detection. RAXO builds high-quality X-ray class descriptors using a dual-source retrieval strategy. It gathers relevant RGB images from the web and enriches them via a novel X-ray material transfer mechanism, eliminating the need for labeled databases. These visual descriptors replace text-based classification in OvOD, leveraging intra-modal feature distances for robust detection. Extensive experiments demonstrate that RAXO consistently improves OvOD performance, providing an average mAP increase of up to 17.0 points over base detectors. To further support research in this emerging field, we also introduce DET-COMPASS, a new benchmark featuring bounding box annotations for over 300 object categories, enabling large-scale evaluation of OvOD in X-ray.
@inproceedings{garcia2025superpowering, title = {Superpowering Open-Vocabulary Object Detectors for X-ray Vision}, author = {Garcia-Fernandez, Pablo and Vaquero, Lorenzo and Liu, Mingxuan and Xue, Feng and Cores, Daniel and Sebe, Nicu and Mucientes, Manuel and Ricci, Elisa}, booktitle = {IEEE International Conference on Computer Vision (ICCV)}, year = {2025}, }