@inproceedings{SerranoMesa2026,
title = {Mapeo semántico de vocabulario abierto mediante 3D Gaussian Splatting},
author = {Serrano-Mesa, Jorge and Moncada-Ramirez, Jesús and Ojeda, Pepe and Ambrosio, Gregorio and Ruiz-Sarmiento, J. R. and Gonzalez-Jimenez, Javier},
url = {https://revistas.udc.es/index.php/JA_CEA/article/view/13803/9700},
doi = {10.17979/ja-cea.2026.47.13803},
year = {2026},
date = {2026-09-01},
urldate = {2026-09-01},
booktitle = {Jornadas de Automática CEA, Córdoba},
journal = {XLVII Jornadas de Automática, Córdoba, Visión por Computador},
number = {47},
abstract = {<jats:p>Open-vocabulary semantic maps are key to the autonomy of mobile robots in complex, unstructured environments. Compared to traditional techniques, open-vocabulary semantic maps represent a significant improvement, as they are not limited by a predefined set of categories. For their construction, geometric representations based on 3D Gaussian Splatting enriched with low-level semantic descriptors have gained popularity. However, robotic operations demand high-level labels, such as textual descriptions and categories. This work presents an evolution of OpenSplat3D that integrates Large Vision-Language Models (LVLMs) for the description and categorization of object instances, whose consistency is validated within the latent space of Vision-Language Models (VLMs). Additionally, the geometric reconstruction has been optimized through dense initialization and the use of a depth-based loss function. Both proposals are validated using the Replica and ScanNet datasets, yielding accurate, semantically enriched maps.</jats:p>},
keywords = {},
pubstate = {published},
tppubtype = {inproceedings}
}
