no code implementations • 22 Apr 2024 • Sophia Sirko-Galouchenko, Alexandre Boulch, Spyros Gidaris, Andrei Bursuc, Antonin Vobecky, Patrick Pérez, Renaud Marlet
Models pretrained with our method exhibit improved BEV semantic segmentation performance, particularly in low-data scenarios.
1 code implementation • 26 Oct 2023 • Gilles Puy, Spyros Gidaris, Alexandre Boulch, Oriane Siméoni, Corentin Sautier, Patrick Pérez, Andrei Bursuc, Renaud Marlet
In particular, thanks to our scalable distillation method named ScaLR, we show that scaling the 2D and 3D backbones and pretraining on diverse datasets leads to a substantial improvement of the feature quality.
1 code implementation • 26 Oct 2023 • Corentin Sautier, Gilles Puy, Alexandre Boulch, Renaud Marlet, Vincent Lepetit
We present a surprisingly simple and efficient method for self-supervision of 3D backbone on automotive Lidar point clouds.
1 code implementation • 6 Apr 2023 • Bjoern Michele, Alexandre Boulch, Gilles Puy, Tuan-Hung Vu, Renaud Marlet, Nicolas Courty
Learning models on one labeled dataset that generalize well on another domain is a difficult task, as several shifts might happen between the data domains.
1 code implementation • ICCV 2023 • Gilles Puy, Alexandre Boulch, Renaud Marlet
Semantic segmentation of point clouds in autonomous driving datasets requires techniques that can process large numbers of points efficiently.
Ranked #4 on LIDAR Semantic Segmentation on nuScenes (val mIoU metric)
1 code implementation • CVPR 2023 • Angelika Ando, Spyros Gidaris, Andrei Bursuc, Gilles Puy, Alexandre Boulch, Renaud Marlet
(c) We refine pixel-wise predictions with a convolutional decoder and a skip connection from the convolutional stem to combine low-level but fine-grained features of the the convolutional stem with the high-level but coarse predictions of the ViT encoder.
1 code implementation • CVPR 2023 • Alexandre Boulch, Corentin Sautier, Björn Michele, Gilles Puy, Renaud Marlet
The core idea is to train the model on a pretext task which is the reconstruction of the surface on which the 3D points are sampled, and to use the underlying latent vectors as input to the perception head.
no code implementations • 14 Apr 2022 • Wei-Hsin Tseng, Hoàng-Ân Lê, Alexandre Boulch, Sébastien Lefèvre, Dirk Tiede
It is of interest to localize a ground-based LiDAR point cloud on remote sensing imagery.
1 code implementation • CVPR 2022 • Corentin Sautier, Gilles Puy, Spyros Gidaris, Alexandre Boulch, Andrei Bursuc, Renaud Marlet
In this context, we propose a self-supervised pre-training method for 3D perception models that is tailored to autonomous driving data.
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 • CVPR 2022 • Alexandre Boulch, Renaud Marlet
To overcome this limitation, a few approaches infer latent vectors on a coarse regular 3D grid or on 3D patches, and interpolate them to answer occupancy queries.
Ranked #2 on 3D Reconstruction on ShapeNet
no code implementations • 31 Dec 2021 • Rodrigo Caye Daudt, Bertrand Le Saux, Alexandre Boulch, Yann Gousseau
Large scale datasets created from crowdsourced labels or openly available data have become crucial to provide training data for large scale learning algorithms.
1 code implementation • 30 Nov 2021 • Alexandre Boulch, Pierre-Alain Langlois, Gilles Puy, Renaud Marlet
There has been recently a growing interest for implicit shape representations.
1 code implementation • ICCV 2021 • Anh-Quan Cao, Gilles Puy, Alexandre Boulch, Renaud Marlet
Rigid registration of point clouds with partial overlaps is a longstanding problem usually solved in two steps: (a) finding correspondences between the point clouds; (b) filtering these correspondences to keep only the most reliable ones to estimate the transformation.
1 code implementation • 13 Aug 2021 • Björn Michele, Alexandre Boulch, Gilles Puy, Maxime Bucher, Renaud Marlet
While there has been a number of studies on Zero-Shot Learning (ZSL) for 2D images, its application to 3D data is still recent and scarce, with just a few methods limited to classification.
Ranked #1 on Generalized Zero-Shot Learning on ScanNet
no code implementations • ICLR Workshop EBM 2021 • Javiera Castillo Navarro, Bertrand Le Saux, Alexandre Boulch, Sébastien Lefèvre
The large amount of data, available thanks to the recent sensors, have made possible the use of deep learning for Earth Observation.
no code implementations • 15 Oct 2020 • Javiera Castillo-Navarro, Bertrand Le Saux, Alexandre Boulch, Nicolas Audebert, Sébastien Lefèvre
The development of semi-supervised learning techniques is essential to enhance the generalization capacities of machine learning algorithms.
1 code implementation • ECCV 2020 • Gilles Puy, Alexandre Boulch, Renaud Marlet
Our main finding is that FLOT can perform as well as the best existing methods on synthetic and real-world datasets while requiring much less parameters and without using multiscale analysis.
1 code implementation • 10 Jul 2020 • Pierre Godet, Alexandre Boulch, Aurélien Plyer, Guy Le Besnerais
We present a new lightweight CNN-based algorithm for multi-frame optical flow estimation.
1 code implementation • 9 Apr 2020 • Alexandre Boulch, Gilles Puy, Renaud Marlet
Recent state-of-the-art methods for point cloud processing are based on the notion of point convolution, for which several approaches have been proposed.
Ranked #1 on LIDAR Semantic Segmentation on Paris-Lille-3D
no code implementations • 4 Nov 2019 • Marcela Carvalho, Maxime Ferrera, Alexandre Boulch, Julien Moras, Bertrand Le Saux, Pauline Trouvé-Peloux
This paper is a technical report about our submission for the ECCV 2018 3DRMS Workshop Challenge on Semantic 3D Reconstruction \cite{Tylecek2018rms}.
2 code implementations • 1 Nov 2019 • Pierre-Alain Langlois, Alexandre Boulch, Renaud Marlet
In man-made environments such as indoor scenes, when point-based 3D reconstruction fails due to the lack of texture, lines can still be detected and used to support surfaces.
no code implementations • 4 Sep 2019 • Nicolas Audebert, Alexandre Boulch, Bertrand Le Saux, Sébastien Lefèvre
This work introduces a new semantic segmentation regularization based on the regression of a distance transform.
no code implementations • 17 Apr 2019 • Rodrigo Caye Daudt, Bertrand Le Saux, Alexandre Boulch, Yann Gousseau
Large scale datasets created from user labels or openly available data have become crucial to provide training data for large scale learning algorithms.
1 code implementation • 4 Apr 2019 • Alexandre Boulch
Point clouds are unstructured and unordered data, as opposed to images.
Ranked #4 on LIDAR Semantic Segmentation on Paris-Lille-3D
1 code implementation • 31 Oct 2018 • Alexandre Boulch, Noëlie Cherrier, Thibaut Castaings
The ionosphere electromagnetic activity is a major factor of the quality of satellite telecommunications, Global Navigation Satellite Systems (GNSS) and other vital space applications.
4 code implementations • 19 Oct 2018 • Rodrigo Caye Daudt, Bertrand Le Saux, Alexandre Boulch
This paper presents three fully convolutional neural network architectures which perform change detection using a pair of coregistered images.
Ranked #1 on Change Detection on OSCD - 13ch
Change Detection Change detection for remote sensing images +1
no code implementations • 19 Oct 2018 • Rodrigo Caye Daudt, Bertrand Le Saux, Alexandre Boulch, Yann Gousseau
In this paper we present the first large scale high resolution semantic change detection (HRSCD) dataset, which enables the usage of deep learning methods for semantic change detection.
1 code implementation • 19 Oct 2018 • Rodrigo Caye Daudt, Bertrand Le Saux, Alexandre Boulch, Yann Gousseau
The Copernicus Sentinel-2 program now provides multispectral images at a global scale with a high revisit rate.
2 code implementations • 28 Feb 2017 • Alexandre Boulch
We show, on the one hand, that they are almost as efficient as their sequential counterparts while involving less parameters, and on the other hand that they are more efficient than a residual network with the same number of parameters.