Search Results for author: Jean-Emmanuel Deschaud

Found 19 papers, 8 papers with code

MDT3D: Multi-Dataset Training for LiDAR 3D Object Detection Generalization

no code implementations2 Aug 2023 Louis Soum-Fontez, Jean-Emmanuel Deschaud, François Goulette

Therefore, we decided to leverage the information available from several annotated source datasets with our Multi-Dataset Training for 3D Object Detection (MDT3D) method to increase the robustness of 3D object detection models when tested in a new environment with a different sensor configuration.

3D Object Detection Domain Adaptation +2

Domain generalization of 3D semantic segmentation in autonomous driving

1 code implementation ICCV 2023 Jules Sanchez, Jean-Emmanuel Deschaud, Francois Goulette

This method relies on leveraging the geometry and sequentiality of the LiDAR data to enhance its generalization performances by working on partially accumulated point clouds.

3D Semantic Segmentation Autonomous Driving +1

Unsigned Distance Field as an Accurate 3D Scene Representation for Neural Scene Completion

no code implementations17 Mar 2022 Jean Pierre Richa, Jean-Emmanuel Deschaud, François Goulette, Nicolas Dalmasso

The proposed UDF is simple, and efficient as a geometry representation, and can be computed on any point cloud.

COLA: COarse LAbel pre-training for 3D semantic segmentation of sparse LiDAR datasets

1 code implementation14 Feb 2022 Jules Sanchez, Jean-Emmanuel Deschaud, François Goulette

Transfer learning is a proven technique in 2D computer vision to leverage the large amount of data available and achieve high performance with datasets limited in size due to the cost of acquisition or annotation.

Autonomous Driving CoLA +3

Paris-CARLA-3D: A Real and Synthetic Outdoor Point Cloud Dataset for Challenging Tasks in 3D Mapping

no code implementations22 Nov 2021 Jean-Emmanuel Deschaud, David Duque, Jean Pierre Richa, Santiago Velasco-Forero, Beatriz Marcotegui, and François Goulette

The data are composed of two sets with synthetic data from the open source CARLA simulator (700 million points) and real data acquired in the city of Paris (60 million points), hence the name Paris-CARLA-3D.

Instance Segmentation Semantic Segmentation

Riedones3D: a celtic coin dataset for registration and fine-grained clustering

no code implementations30 Sep 2021 Sofiane Horache, Jean-Emmanuel Deschaud, François Goulette, Katherine Gruel, Thierry Lejars, Olivier Masson

With this dataset, we propose two benchmarks, one for point cloud registration, essential for coin die recognition, and a benchmark of coin die clustering.

Clustering Cultural Vocal Bursts Intensity Prediction +1

CT-ICP: Real-time Elastic LiDAR Odometry with Loop Closure

1 code implementation27 Sep 2021 Pierre Dellenbach, Jean-Emmanuel Deschaud, Bastien Jacquet, François Goulette

Multi-beam LiDAR sensors are increasingly used in robotics, particularly with autonomous cars for localization and perception tasks, both relying on the ability to build a precise map of the environment.

Loop Closure Detection RTE

KITTI-CARLA: a KITTI-like dataset generated by CARLA Simulator

1 code implementation17 Aug 2021 Jean-Emmanuel Deschaud

This dataset thus makes it possible to improve transfer learning methods from a synthetic dataset to a real dataset.

Semantic Segmentation Transfer Learning

What's in My LiDAR Odometry Toolbox?

1 code implementation17 Mar 2021 Pierre Dellenbach, Jean-Emmanuel Deschaud, Bastien Jacquet, François Goulette

With the democratization of 3D LiDAR sensors, precise LiDAR odometries and SLAM are in high demand.

Automatic clustering of Celtic coins based on 3D point cloud pattern analysis

no code implementations12 May 2020 Sofiane Horache, Jean-Emmanuel Deschaud, François Goulette, Katherine Gruel, Thierry Lejars

The recognition and clustering of coins which have been struck by the same die is of interest for archeological studies.

Clustering

IMLS-SLAM: scan-to-model matching based on 3D data

no code implementations23 Feb 2018 Jean-Emmanuel Deschaud

The Simultaneous Localization And Mapping (SLAM) problem has been well studied in the robotics community, especially using mono, stereo cameras or depth sensors.

Robotics

Cannot find the paper you are looking for? You can Submit a new open access paper.