From ea8e9a0878ee9500f916b6fb6610d1a92290a2cf Mon Sep 17 00:00:00 2001 From: "d.pascualhe" Date: Fri, 15 May 2026 10:50:18 +0200 Subject: [PATCH] Add new publications, restructure GAIA videos, and fix format --- _pages/projects/gaia.md | 111 ++++++++++-------- ..._models_for_off_road_autonomous_driving.md | 68 +++++++++++ _pages/publications/publications.md | 34 ++++-- 3 files changed, 154 insertions(+), 59 deletions(-) create mode 100644 _pages/publications/2026/cross_dataset_evaluation_of_visual_semantic_segmentation_models_for_off_road_autonomous_driving.md diff --git a/_pages/projects/gaia.md b/_pages/projects/gaia.md index b6231eb..9c38d44 100644 --- a/_pages/projects/gaia.md +++ b/_pages/projects/gaia.md @@ -38,24 +38,24 @@ classes: wide height: 100%; object-fit: cover; } - +
-
+
- +
-
+
- +
-
- +
+
@@ -69,29 +69,29 @@ Wildfires that burn for weeks and affect millions of people represent a challeng The main 3 contributions of the RoboticsLab group are: - * LIDAR Signal Densification Algorithms: Research focuses on algorithms and software techniques for densifying the LIDAR sensor signal, commonly used in outdoor robots. + * LIDAR Signal Densification Algorithms: Research focuses on algorithms and software techniques for densifying the LIDAR sensor signal, commonly used in outdoor robots. - * Outdoor Navigation Algorithms: This involves developing and evaluating point-to-point autonomous navigation algorithms for robots in unstructured environments, such as forests. + * Outdoor Navigation Algorithms: This involves developing and evaluating point-to-point autonomous navigation algorithms for robots in unstructured environments, such as forests. - * Deep Learning for Object and Surface Detection: Evolution of techniques for detecting objects and surfaces in unstructured environments using Deep Learning and sensors like LIDAR and visual cameras, aimed at autonomous robot navigation. + * Deep Learning for Object and Surface Detection: Evolution of techniques for detecting objects and surfaces in unstructured environments using Deep Learning and sensors like LIDAR and visual cameras, aimed at autonomous robot navigation.
-
- +
+
## Videos -### Lidar Segmentation +### LiDAR Segmentation
+
+
+
+
+
+ +### Simulated Environments +
+
-
+
@@ -223,7 +232,7 @@ The main 3 contributions of the RoboticsLab group are:
-
+

Realistic Unstructured environment v3

- - + +
-
+

Unstructured environment v2

-
+

Unstructured environment

@@ -347,6 +356,18 @@ The main 3 contributions of the RoboticsLab group are: ## Publications +
+
+
Cross-Dataset Evaluation of Visual Semantic Segmentation Models for Off-Road Autonomous Driving
+
David Pascual-Hernández, Sergio Paniego, Roberto Calvo-Palomino, Inmaculada Mora-Jiménez, Jose Maria Cañas-Plaza
+
Expert Systems with Applications, Elsevier, 2026
+
+
+
Deep Learning-Based Semantic Segmentation of LiDAR Point Clouds for Autonomous Driving in Unstructured Environments
+
Félix Martínez, David Pascual-Hernández, Daniel Borja Fernández, Inmaculada Mora Jiménez, José María Cañas
+
Proceedings of the XXV International Workshop on Physical Agents (WAF), 2025
+
+
@@ -374,21 +395,19 @@ The main 3 contributions of the RoboticsLab group are:
SoftwareX, Elsevier, 2024
- +
## Founded by - +
- Paper + Paper
- -
- +
diff --git a/_pages/publications/2026/cross_dataset_evaluation_of_visual_semantic_segmentation_models_for_off_road_autonomous_driving.md b/_pages/publications/2026/cross_dataset_evaluation_of_visual_semantic_segmentation_models_for_off_road_autonomous_driving.md new file mode 100644 index 0000000..3359cc5 --- /dev/null +++ b/_pages/publications/2026/cross_dataset_evaluation_of_visual_semantic_segmentation_models_for_off_road_autonomous_driving.md @@ -0,0 +1,68 @@ +--- +permalink: /publications/2026/cross_dataset_evaluation_of_visual_semantic_segmentation_models_for_off_road_autonomous_driving +title: "Cross-dataset Evaluation of Visual Semantic Segmentation Models for Off-road Autonomous Driving" + + +layout: single + +classes: wide +--- + +

Neurocomputing, 2024

+ +

David Pascual-Hernández1, Sergio Paniego1, Roberto Calvo-Palomino1

, Inmaculada Mora-Jiménez1

, Jose Maria Cañas-Plaza1

+
+

1: URJC

+
+

DOI: 10.1016/j.eswa.2026.132656

+ + +## Abstract + +Intelligent autonomous driving in off-road environments is an emerging field with great potential to impact areas such as agriculture, forestry, and rescue operations. Perception in these scenarios presents unique challenges due to the diversity of elements and weather conditions, along with the inherent ambiguity in class definitions. Consequently, off-road visual semantic segmentation datasets remain underdeveloped, roughly ten times smaller than their urban counterparts, hindering dependable performance assessment and potentially compromising the safety of autonomous systems. To address these challenges, we present a comprehensive cross-dataset evaluation of visual semantic segmentation models for autonomous off-road navigation. We propose a unified ontology that harmonizes class definitions across relevant datasets, enabling their combination for both training and testing. This approach ensures fair model comparisons and reliable assessment of generalization to unseen domains. We further benchmark models on the original datasets, analyze the impact of different ontology harmonization criteria and conversion strategies, and evaluate the trade-off between segmentation performance and computational cost. Results show that Transformer-based architectures achieve the most consistent segmentation performance across datasets. While often computationally demanding, some variants maintain real-time inference (≈12 ms) with top-tier accuracy. The unified ontology simplifies the segmentation task, yielding more reliable models and about 40% faster training convergence. Cross-dataset training further enhances generalization, improving mean IoU by up to +20% on RUGD and +13% on WildScenes compared to RELLIS-3D-only training. Overall, this study provides valuable insights for developing robust perception modules for off-road autonomous vehicles. + +
+
+ +
+
+ +
+
+ +## Materials + +
+ + + +
+ + +## Citation + +``` +@article{pascual2026cross, + title={Cross-Dataset Evaluation of Visual Semantic Segmentation Models for Off-Road Autonomous Driving}, + author={Pascual-Hern{\'a}ndez, David and Paniego, Sergio and Calvo-Palomino, Roberto and Mora-Jim{\'e}nez, Inmaculada and Ca{\~n}as-Plaza, Jos{\'e} Mar{\'\i}a}, + journal={Expert Systems with Applications}, + pages={132656}, + year={2026}, + publisher={Elsevier} +} +``` diff --git a/_pages/publications/publications.md b/_pages/publications/publications.md index ad9d6fd..9042f64 100644 --- a/_pages/publications/publications.md +++ b/_pages/publications/publications.md @@ -12,6 +12,14 @@ classes: wide Journals, congress papers and research publications can be found below: +# 2026 + +* [Cross-Dataset Evaluation of Visual Semantic Segmentation Models for Off-Road Autonomous Driving, Expert Systems with Applications. David Pascual-Hernández; Sergio Paniego; Roberto Calvo-Palomino; Inmaculada Mora-Jiménez; Jose Maria Cañas-Plaza](/publications/2026/cross_dataset_evaluation_of_visual_semantic_segmentation_models_for_off_road_autonomous_driving). + +# 2025 + +* [Deep Learning-Based Semantic Segmentation of LiDAR Point Clouds for Autonomous Driving in Unstructured Environments, Proceedings of the XXV International Workshop on Physical Agents (WAF). Félix Martínez, David Pascual-Hernández, Daniel Borja Fernández, Inmaculada Mora Jiménez, José María Cañas] + # 2024 * [Enhancing End-to-End Control in Autonomous Driving through Kinematic-Infused and Visual Memory Imitation Learning Neurocomputing. Sergio Paniego; Roberto Calvo-Palomino; JoséMaría Cañas](/publications/2024/enhancing_end_to_end_control_in_autonomous_driving_through_kinematic_infused_and_visual_memory_imitation_learning). @@ -23,10 +31,10 @@ Journals, congress papers and research publications can be found below: * [Model Optimization in Deep Learning based Robot Control for Autonomous Driving IEEE Robotics and Automation Letters (RA-L) & ICRA. Sergio Paniego; Nikhil Paliwal; JoseMaria Cañas](/publications/2023/model_optimization_in_deep_learning_based_robot_control_for_autonomous_driving). -* Experimental analysis of the effectiveness of a Cyber-physical Robotic System to assist speech and language pathologists in high school. Journal of New Approaches in Educational Research. U. Alicante. Eldon Caldwell Marín; Miguel Cazorla; José María Cañas. +* Experimental analysis of the effectiveness of a Cyber-physical Robotic System to assist speech and language pathologists in high school. Journal of New Approaches in Educational Research. U. Alicante. Eldon Caldwell Marín; Miguel Cazorla; José María Cañas. * A middleware infrastructure for programming vision-based applications in UAVs. Drones. MDPI. 6-11. Pedro Arias Pérez; Jesús Fernández Conde; David Martín Gómez; José -María Cañas Plaza; Pascual Campoy. +María Cañas Plaza; Pascual Campoy. * [Open Source Assessment of Deep Learning Visual Object Detection. Sensors. MDPI. 22-12. Sergio Paniego Blanco; Vinay Sharma; José María Cañas Plaza](/publications/2022/open_source_assessment_of_deep_learning_visual_object_detection). @@ -34,7 +42,7 @@ María Cañas Plaza; Pascual Campoy. # 2022 -* Efficient 3D Human Pose Estimation from RGBD sensors. DISPLAYS. Elsevier. David Pascual-Hernández; Nuria Oyaga de Frutos; Inmaculada Mora Jiménez. +* Efficient 3D Human Pose Estimation from RGBD sensors. DISPLAYS. Elsevier. David Pascual-Hernández; Nuria Oyaga de Frutos; Inmaculada Mora Jiménez. * Improving Performance of Temporal Difference Learning in Board Games. Applied Sciences. MPDI. 6-12. Jesús Fernández-Conde; Pedro-Manuel Cuenca-Jiménez; José M. Cañas Plaza. 2022. Hybrid Training Strategies @@ -48,15 +56,15 @@ María Cañas Plaza; Pascual Campoy. * Computing for Reactive Robotics Using Open-Source FPGAs. José María Cañas Plaza; Jesús Fernández Conde; Julio Vega Pérez; Juan Ordoñez Cerezo. 2021. Reconfigurable Electronics. MDPI. 11-1. * Designing a Cyber-physical Robotic Platform to assist Speech-Language Pathologists. Assistive Technology. Taylor & -Francis. Eldon Glen Caldwell Marin; Carlos Andres Morales; Emilia Solís Sánchez; Miguel Cazorla; José María Cañas Plaza. 2021. +Francis. Eldon Glen Caldwell Marin; Carlos Andres Morales; Emilia Solís Sánchez; Miguel Cazorla; José María Cañas Plaza. 2021. * Adaptive Uplink Data Compression in Spectrum Crowdsensing Systems Y Zeng, R Calvo-Palomino, D Giustiniano, G Bovet, S Banerjee. 2021 IEEE International Symposium on Dynamic Spectrum Access Networkw # 2020 -* Open-Source Drone Programming Course for Distance Engineering Education. Electronics. MDPI. 9-12. José María; Diego; Pedro; Julio; David; Lía; Jesús. 2020. +* Open-Source Drone Programming Course for Distance Engineering Education. Electronics. MDPI. 9-12. José María; Diego; Pedro; Julio; David; Lía; Jesús. 2020. -* A ROS‐Based Open Tool for Intelligent Robotics Education. Applied Sciences. MDPI. 10-21. José María; Eduardo; Lía; Jesús. 2020. +* A ROS‐Based Open Tool for Intelligent Robotics Education. Applied Sciences. MDPI. 10-21. José María; Eduardo; Lía; Jesús. 2020. * LSTM-based GNSS Spoofing Detection Using Low-cost Spectrum Sensors. R Calvo-Palomino, A Bhattacharya, G Bovet, D Giustiniano. The 21st IEEE International Symposium on a World of Wireless, Mobile 2020 @@ -64,24 +72,24 @@ Francis. Eldon Glen Caldwell Marin; Carlos Andres Morales; Emilia Solís # 2019 -* Open Vision System for Low-Cost Robotics Education. Electronics. MDPI. 8-11, pp.1295. Artículo científico. Julio Vega Pérez. 2019. +* Open Vision System for Low-Cost Robotics Education. Electronics. MDPI. 8-11, pp.1295. Artículo científico. Julio Vega Pérez. 2019. * COMBAHO: A deep learning system for integrating brain injury patients in society. Pattern Recognition Letters. Elsevier. Garcia-Rodriguez, Jose; Gomez-Donoso, Francisco; Oprea, Sergiu; et -al; others. 2019. +al; others. 2019. * Control system in open-source FPGA for a self-balancing -robot. Electronics. Multidisciplinary Digital Publishing Institute. 8-2, pp.198-198. Ordoñez Cerezo, Juan; Castillo Morales, Encarnaci{\'o}n; Canas Plaza, Jose Maria. 2019. +robot. Electronics. Multidisciplinary Digital Publishing Institute. 8-2, pp.198-198. Ordoñez Cerezo, Juan; Castillo Morales, Encarnaci{\'o}n; Canas Plaza, Jose Maria. 2019. * Enhancing the ambient assisted living capabilities with a mobile robot. Computational intelligence and neuroscience. Hindawi. 2019. Gomez-Donoso, Francisco; Escalona, F{\'e}lix; Rivas, Francisco -Miguel; Ca{\~n}as, Jose Maria; Cazorla, Miguel. 2019. +Miguel; Ca{\~n}as, Jose Maria; Cazorla, Miguel. 2019. * PyBoKids: An Innovative Python-Based Educational Framework Using Real and Simulated Arduino -Robots. Electronics. Multidisciplinary Digital Publishing Institute. 8-8, pp.899-899. Vega, Julio; Cañas, José M. 2019. +Robots. Electronics. Multidisciplinary Digital Publishing Institute. 8-8, pp.899-899. Vega, Julio; Cañas, José M. 2019. * SDVL: Efficient and accurate semi-direct visual localization. Sensors. Multidisciplinary Digital Publishing -Institute. 19-2, pp.302-302. Perdices, Eduardo; Cañas, José María. 2019. +Institute. 19-2, pp.302-302. Perdices, Eduardo; Cañas, José María. 2019. * Quantitative analysis of security in distributed robotic frameworks. Robotics and Autonomous Systems. 100, pp.95. ISSN 0921-8890. Martín, Francisco; Soriano, Enrique; Cañas, José María. (/3). 2018. https://doi.org/10.1016/j.robot.2017.11.002 @@ -97,4 +105,4 @@ Rivas, Francisco; Calvo, Roberto. 2018. Revista Iberoamericana de Automática e using a single camera. Kachach, Redouane; Cañas, José María. (/2). 2016. Journal of Electronic Imaging. 25/3. ISSN 1017-9909. https://doi.org/10.1117/1.JEI.25.3.033021 -* Social robots in advanced dementia. Frontiers in Aging Neuroscience. Valenti, Meritxell; Cañas, José María. (/4). 2015. +* Social robots in advanced dementia. Frontiers in Aging Neuroscience. Valenti, Meritxell; Cañas, José María. (/4). 2015.