LATIDIA · Robótica

Hacia la ampliación de la percepción marina con datos sintéticos

arXiv:2609.20680v1 Tipo de anuncio: nuevo Resumen: El aprendizaje automático escalable en entornos submarinos desafiantes está fuertemente limitado por la falta de datos de entrenamiento etiquetados del mundo real. Estos datos suelen ser caros y laboriosos

La noticia

arXiv:2609.20680v1 Announce Type: new Abstract: Scalable machine learning in challenging underwater environments is strongly limited by the lack of labeled real-world training data. This data is often expensive and laborious to gather, making large-scale real-world data challenging to gather and curate. However, simulated data can help close the gap, enabling many learning-based tasks for underwater perception. In this work, we extend OceanSim, an IsaacSim-based underwater perception simulator, with a Synthetic Data Generation (SDG) pipeline for training models to be used in underwater scenarios. The proposed pipeline enables users to generate large, automatically labeled, photorealistic datasets with configurable scene appearance, structure, and sensor settings. We evaluate the pipeline on a real-world