Publications de Mathieu Salzmann

Publications associées (159)

Eigendecomposition-Free Training of Deep Networks for Linear Least-Square Problems

Pascal Fua, Mathieu Salzmann, Kwang Moo Yi, Yinlin Hu, Zheng Dang

Many classical Computer Vision problems, such as essential matrix computation and pose estimation from 3D to 2D correspondences, can be tackled by solving a linear least-square problem, which can be done by finding the eigenvector corresponding to the smal ...

IEEE COMPUTER SOC2021

Learning and recovering 3D surface deformations

Mathieu Salzmann

Recovering the 3D deformations of a non-rigid surface from a single viewpoint has applications in many domains such as sports, entertainment, and medical imaging. Unfortunately, without any knowledge of the possible deformations that the object of interest ...

EPFL2009

GigaPose: Fast and Robust Novel Object Pose Estimation via One Correspondence

Mathieu Salzmann

We present GigaPose, a fast, robust, and accurate method for CAD-based novel object pose estimation in RGB images. GigaPose first leverages discriminative "templates", rendered images of the CAD models, to recover the out-of-plane rotation and then uses pa ...

IEEE2024

Automatic Content Curation of Visual Heritage

Mathieu Salzmann, Frédéric Kaplan, Nicolas Henchoz, Delphine Ribes Lemay, Lara Jacqueline Défayes, André Viegas Andrade

Digitization and preservation of large heritage induce high maintenance costs to keep up with the technical standards and ensure sustainable access. Creating impactful usage is instrumental to justify the resources for long-term preservation. The Museum fü ...

2021

Learning to Segment 3D Linear Structures Using Only 2D Annotations

Pascal Fua, Mathieu Salzmann, Mateusz Kozinski, Agata Justyna Mosinska

We propose a loss function for training a Deep Neural Network (DNN) to segment volumetric data, that accommodates ground truth annotations of 2D projections of the training volumes, instead of annotations of the 3D volumes themselves. In consequence, we si ...

SPRINGER INTERNATIONAL PUBLISHING AG2018

Attention-based domain adaptation for single-stage detectors

Mathieu Salzmann, Vidit Vidit

While domain adaptation has been used to improve the performance of object detectors when the training and test data follow different distributions, previous work has mostly focused on two-stage detectors. This is because their use of region proposals make ...

SPRINGER2022

Counting People by Estimating People Flows

Pascal Fua, Mathieu Salzmann, Weizhe Liu

Modern methods for counting people in crowded scenes rely on deep networks to estimate people densities in individual images. As such, only very few take advantage of temporal consistency in video sequences, and those that do only impose weak smoothness co ...

2022

Weakly-supervised Action Transition Learning for Stochastic Human Motion Prediction

Mathieu Salzmann

We introduce the task of action-driven stochastic human motion prediction, which aims to predict multiple plausible future motions given a sequence of action labels and a short motion history. This differs from existing works, which predict motions that ei ...

IEEE COMPUTER SOC2022

Perspective Aware Road Obstacle Detection

Pascal Fua, Mathieu Salzmann, Sina Honari, Krzysztof Maciej Lis

While road obstacle detection techniques have become increasingly effective, they typically ignore the fact that, in practice, the apparent size of the obstacles decreases as their distance to the vehicle increases. In this letter, we account for this by c ...

2023

Training Provably Robust Models by Polyhedral Envelope Regularization

Sabine Süsstrunk, Mathieu Salzmann, Chen Liu

Training certifiable neural networks enables us to obtain models with robustness guarantees against adversarial attacks. In this work, we introduce a framework to obtain a provable adversarial-free region in the neighborhood of the input data by a polyhedr ...

IEEE-INST ELECTRICAL ELECTRONICS ENGINEERS INC2023

Deep Learning to Improve the Discovery of Near-earth Asteroids in the Zwicky Transient Facility

Belén Yu Irureta-Goyena Chang, Jean-Paul Richard Kneib, Mathieu Salzmann, Frédéric Dux

We present a novel pipeline that uses a convolutional neural network (CNN) to improve the detection capability of near-Earth asteroids (NEAs) in the context of planetary defense. Our work aims to minimize the dependency on human intervention of the current ...

IOP Publishing Ltd2025

Using Motion Cues to Supervise Single-Frame Body Pose and Shape Estimation in Low Data Regimes

Pascal Fua, Mathieu Salzmann, Andrey Davydov

When enough annotated training data is available, supervised deep-learning. algorithms excel at estimating human body pose and shape using a single camera. The effects of too little such data being available can be mitigated by using other information sour ...

Journal of Machine Learning Research Inc.2024

Joint Compression of Multi-model Systems for Edge Devices

Mathieu Salzmann, João Luís Prado Vieira

Deploying deep learning models for computer vision (CV) often requires a careful trade-off between performance, efficiency, and cost. In the context of edge inference of machine learning systems, this trade-off translates into a challenge, as computing and ...

Springer Science and Business Media Deutschland GmbH2025

Memory Efficient Max Flow for Multi-Label Submodular MRFs

Mathieu Salzmann

Multi-label submodular Markov Random Fields (MRFs) have been shown to be solvable using max-flow based on an encoding of the labels proposed by Ishikawa, in which each variable X-i is represented by l nodes (where l is the number of labels) arranged in a c ...

2019

SD-Pose: Semantic Decomposition for Cross-Domain 6D Object Pose Estimation

Mathieu Salzmann, Yinlin Hu

The current leading 6D object pose estimation methods rely heavily on annotated real data, which is highly costly to acquire. To overcome this, many works have proposed to introduce computer-generated synthetic data. However, bridging the gap between the s ...

ASSOC ADVANCEMENT ARTIFICIAL INTELLIGENCE2021

Distribution-Matching Embedding for Visual Domain Adaptation

Mathieu Salzmann

Domain-invariant representations are key to addressing the domain shift problem where the training and test examples follow different distributions. Existing techniques that have attempted to match the distributions of the source and target domains typical ...

2016

Learning to Generate the Unknowns as a Remedy to the Open-Set Domain Shift

Mathieu Salzmann

In many situations, the data one has access to at test time follows a different distribution from the training data. Over the years, this problem has been tackled by closed-set domain adaptation techniques. Recently, open-set domain adaptation has emerged ...

IEEE COMPUTER SOC2022

Temporally-Consistent Surface Reconstruction Using Metrically-Consistent Atlases

Pascal Fua, Mathieu Salzmann, Shaifali Parashar, Jan Bednarík

We propose a method for unsupervised reconstruction of a temporally-consistent sequence of surfaces from a sequence of time-evolving point clouds. It yields dense and semantically meaningful correspondences between frames. We represent the reconstructed su ...

2025

TIC-TAC: A Framework for Improved Covariance Estimation in Deep Heteroscedastic Regression

Mathieu Salzmann, Alexandre Massoud Alahi, Megh Hiren Shukla

Deep heteroscedastic regression involves jointly optimizing the mean and covariance of the predicted distribution using the negative log-likelihood. However, recent works show that this may result in sub-optimal convergence due to the challenges associated ...

2024

Linear Local Models for Monocular Reconstruction of Deformable Surfaces

Pascal Fua, Mathieu Salzmann

Recovering the 3D shape of a nonrigid surface from a single viewpoint is known to be both ambiguous and challenging. Resolving the ambiguities typically requires prior knowledge about the most likely deformations that the surface may undergo. It often take ...

2011