6 code implementations • 9 Oct 2020 • Thomas Chaton, Nicolas Chaulet, Sofiane Horache, Loic Landrieu
We introduce Torch-Points3D, an open-source framework designed to facilitate the use of deep networks on3D data.
2 code implementations • CVPR 2018 • Loic Landrieu, Martin Simonovsky
We propose a novel deep learning-based framework to tackle the challenge of semantic segmentation of large-scale point clouds of millions of points.
Ranked #5 on Semantic Segmentation on Semantic3D
2 code implementations • CVPR 2019 • Loic Landrieu, Mohamed Boussaha
We propose a new supervized learning framework for oversegmenting 3D point clouds into superpoints.
Ranked #28 on Semantic Segmentation on S3DIS
1 code implementation • 10 May 2019 • Loic Landrieu, Mohamed Boussaha
We introduce the graph-structured contrastive loss, a loss function structured by a ground truth segmentation.
1 code implementation • ICCV 2023 • Damien Robert, Hugo Raguet, Loic Landrieu
We introduce a novel superpoint-based transformer architecture for efficient semantic segmentation of large-scale 3D scenes.
Ranked #1 on 3D Semantic Segmentation on S3DIS (mIoU (6-Fold) metric)
1 code implementation • 12 Jan 2024 • Damien Robert, Hugo Raguet, Loic Landrieu
We introduce a highly efficient method for panoptic segmentation of large 3D point clouds by redefining this task as a scalable graph clustering problem.
Ranked #1 on Panoptic Segmentation on DALES
1 code implementation • CVPR 2022 • Damien Robert, Bruno Vallet, Loic Landrieu
Recent works on 3D semantic segmentation propose to exploit the synergy between images and point clouds by processing each modality with a dedicated network and projecting learned 2D features onto 3D points.
Ranked #1 on 3D Semantic Segmentation on KITTI-360 (using extra training data)
2 code implementations • CVPR 2020 • Vivien Sainte Fare Garnot, Loic Landrieu, Sebastien Giordano, Nesrine Chehata
Satellite image time series, bolstered by their growing availability, are at the forefront of an extensive effort towards automated Earth monitoring by international institutions.
Ranked #2 on Time Series Classification on s2-agri
1 code implementation • 14 Dec 2021 • Vivien Sainte Fare Garnot, Loic Landrieu, Nesrine Chehata
Motivated by the recent success of temporal attention-based methods across multiple crop mapping tasks, we propose to investigate how these models can be adapted to operate on several modalities.
Ranked #1 on Panoptic Segmentation on PASTIS-R
2 code implementations • 12 Apr 2024 • Guillaume Astruc, Nicolas Gonthier, Clement Mallet, Loic Landrieu
To demonstrate the advantages of combining modalities of different natures, we augment two existing datasets with new modalities.
1 code implementation • ICCV 2021 • Vivien Sainte Fare Garnot, Loic Landrieu
We also introduce PASTIS, the first open-access SITS dataset with panoptic annotations.
Ranked #3 on Cloud Removal on SEN12MS-CR-TS
1 code implementation • 1 Jul 2020 • Vivien Sainte Fare Garnot, Loic Landrieu
The increasing accessibility and precision of Earth observation satellite data offers considerable opportunities for industrial and state actors alike.
Ranked #1 on Time Series Classification on s2-agri
1 code implementation • 20 Oct 2023 • Anatol Garioud, Nicolas Gonthier, Loic Landrieu, Apolline De Wit, Marion Valette, Marc Poupée, Sébastien Giordano, Boris Wattrelos
We introduce the French Land cover from Aerospace ImageRy (FLAIR), an extensive dataset from the French National Institute of Geographical and Forest Information (IGN) that provides a unique and rich resource for large-scale geospatial analysis.
1 code implementation • 19 Apr 2023 • Romain Loiseau, Elliot Vincent, Mathieu Aubry, Loic Landrieu
We demonstrate the usefulness of our model on a novel dataset of seven large aerial LiDAR scans from diverse real-world scenarios.
1 code implementation • 13 Jul 2021 • Raphael Sulzer, Loic Landrieu, Renaud Marlet, Bruno Vallet
We introduce a novel learning-based, visibility-aware, surface reconstruction method for large-scale, defect-laden point clouds.
1 code implementation • 3 Feb 2022 • Raphael Sulzer, Loic Landrieu, Alexandre Boulch, Renaud Marlet, Bruno Vallet
Most current neural networks for reconstructing surfaces from point clouds ignore sensor poses and only operate on raw point locations.
1 code implementation • 6 Jul 2020 • Vivien Sainte Fare Garnot, Loic Landrieu
In this paper, we propose to model the hierarchical class structure by integrating this metric in the supervision of a prototypical network.
1 code implementation • 31 Jan 2023 • Raphael Sulzer, Renaud Marlet, Bruno Vallet, Loic Landrieu
We present a comprehensive survey and benchmark of both traditional and learning-based methods for surface reconstruction from point clouds.
1 code implementation • 25 Apr 2022 • Ekaterina Kalinicheva, Loic Landrieu, Clément Mallet, Nesrine Chehata
The analysis of the multi-layer structure of wild forests is an important challenge of automated large-scale forestry.
1 code implementation • 15 Oct 2021 • Félix Quinton, Loic Landrieu
While annual crop rotations play a crucial role for agricultural optimization, they have been largely ignored for automated crop type mapping.
no code implementations • 29 Jan 2019 • Vivien Sainte Fare Garnot, Loic Landrieu, Sebastien Giordano, Nesrine Chehata
In this article, we investigate several structured deep learning models for crop type classification on multi-spectral time series.
no code implementations • 27 Dec 2021 • Ekaterina Kalinicheva, Loic Landrieu, Clément Mallet, Nesrine Chehata
We propose a new deep learning-based method for estimating the occupancy of vegetation strata from 3D point clouds captured from an aerial platform.
no code implementations • 20 Jan 2022 • Ekaterina Kalinicheva, Loic Landrieu, Clément Mallet, Nesrine Chehata
We propose a new deep learning-based method for estimating the occupancy of vegetation strata from airborne 3D LiDAR point clouds.
no code implementations • 29 Mar 2024 • Sidi Wu, Yizi Chen, Samuel Mermet, Lorenz Hurni, Konrad Schindler, Nicolas Gonthier, Loic Landrieu
Most image-to-image translation models postulate that a unique correspondence exists between the semantic classes of the source and target domains.