This comparative study evaluates benchmark algorithms for 3D object detection using sensor fusion techniques, highlighting performance in autonomous driving scenarios.
Abstract The use of advanced benchmark algorithms in 3D object detection is essential to improve the safety, efficiency and adaptabilityof autonomous systems operating in complex and changing environments. Numerous benchmark models exist in the scientificcommunity, so it is crucial to identify the most efficient ones. This research work focuses on a comparative study of suchalgorithms based on sensory fusion for autonomous driving. The RGB and LiDAR camera data fusion selected algorithmshave been run on an embedded reference system in order to evaluate their performance and efficiency in a specific 3D objectdetection task. For this purpose, a reference benchmark of the scientific community1 has been used, allowing an objective andstandardised performance evaluation. Finally, as a use case, tests have been run to analyse the robustness and reliability ofthese algorithms against a variety of challenging driving scenarios.
No takes yet. Share an insight, caveat, or question.
Durán et al. (2025) studied this question.
Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context: