Home

Person

Mathieu Salzmann

About
Privacy
Disclaimer

Graph Chatbot

Related publications (169)

Weakly-supervised Action Transition Learning for Stochastic Human Motion Prediction

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

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

Self-supervised Segmentation via Background Inpainting

Pascal Fua, Mathieu Salzmann, Victor Constantin, Isinsu Katircioglu, Helge Jochen Rhodin

While supervised object detection and segmentation methods achieve impressive accuracy, they generalize poorly to images whose appearance significantly differs from the data they have been trained on. To address this when annotating data is prohibitively e ...

2022

An Analysis of Super-Net Heuristics in Weight-Sharing NAS

Mathieu Salzmann, Kaicheng Yu

Weight sharing promises to make neural architecture search (NAS) tractable even on commodity hardware. Existing methods in this space rely on a diverse set of heuristics to design and train the shared-weight backbone network, a.k.a. the super-net. Since he ...

IEEE COMPUTER SOC2022

Progressive Correspondence Pruning by Consensus Learning

Mathieu Salzmann, Hongwei Li, Chenxu Zhao, Feng Zhu, Rui Zhao

Correspondence pruning aims to correctly remove false matches (outliers) from an initial set of putative correspondences. The pruning process is challenging since putative matches are typically extremely unbalanced, largely dominated by outliers, and the r ...

IEEE2021

Human Detection and Segmentation via Multi-view Consensus

Pascal Fua, Mathieu Salzmann, Isinsu Katircioglu, Helge Jochen Rhodin

Self-supervised detection and segmentation of foreground objects aims for accuracy without annotated training data. However, existing approaches predominantly rely on restrictive assumptions on appearance and motion. For scenes with dynamic activities and ...

IEEE2021

Robust Differentiable SVD

Pascal Fua, Mathieu Salzmann, Zheng Dang, Yinlin Hu, Wei Wang

Eigendecomposition of symmetric matrices is at the heart of many computer vision algorithms. However, the derivatives of the eigenvectors tend to be numerically unstable, whether using the SVD to compute them analytically or using the Power Iteration (PI) ...

2021

Geometry-Aware Deep Recurrent Neural Networks for Hyperspectral Image Classification

Mathieu Salzmann, Wei Wang, Siyuan Hao

Variants of deep networks have been widely used for hyperspectral image (HSI)-classification tasks. Among them, in recent years, recurrent neural networks (RNNs) have attracted considerable attention in the remote sensing community. However, complex geomet ...

2021

PCLs: Geometry-aware Neural Reconstruction of 3D Pose with Perspective Crop Layers

Mathieu Salzmann, Helge Jochen Rhodin

Local processing is an essential feature of CNNs and other neural network architectures - it is one of the reasons why they work so well on images where relevant information is, to a large extent, local. However, perspective effects stemming from the proje ...

2021

Temporally-Coherent Surface Reconstruction via Metric-Consistent Atlases

Pascal Fua, Mathieu Salzmann, Siddhartha Chaudhuri, Shaifali Parashar, Jan Bednarík

We propose a method for the unsupervised reconstruction of a temporally-coherent sequence of surfaces from a sequence of time-evolving point clouds, yielding dense, semantically meaningful correspondences between all keyframes. We represent the reconstruct ...

IEEE2021

Towards Robust Fine-grained Recognition by Maximal Separation of Discriminative Features

Mathieu Salzmann

Adversarial attacks have been widely studied for general classification tasks, but remain unexplored in the context of fine-grained recognition, where the inter-class similarities facilitate the attacker's task. In this paper, we identify the proximity of ...

Springer Nature2021

Wide-Depth-Range 6D Object Pose Estimation in Space

Pascal Fua, Mathieu Salzmann, Wenzel Alban Jakob, Yinlin Hu

6D pose estimation in space poses unique challenges that are not commonly encountered in the terrestrial setting. One of the most striking differences is the lack of atmospheric scattering, allowing objects to be visible from a great distance while complic ...

IEEE2021

On the Loss Landscape of Adversarial Training: Identifying Challenges and How to Overcome Them

Mathieu Salzmann, Sabine Süsstrunk, Tao Lin, Chen Liu

We analyze the influence of adversarial training on the loss landscape of machine learning models. To this end, we first provide analytical studies of the properties of adversarial loss functions under different adversarial budgets. We then demonstrate tha ...

2021

Automatic Content Curation of Visual Heritage

Mathieu Salzmann, Nicolas Henchoz, Delphine Ribes Lemay, Frédéric Kaplan, Valentine Bernasconi

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

Contextually Plausible and Diverse 3D Human Motion Prediction

Mathieu Salzmann, Fatemehsadat Saleh

We tackle the task of diverse 3D human motion prediction, that is, forecasting multiple plausible future 3D poses given a sequence of observed 3D poses. In this context, a popular approach consists of using a Conditional Variational Autoencoder (CVAE). How ...

IEEE2021

Probabilistic Tracklet Scoring and Inpainting for Multiple Object Tracking

Mathieu Salzmann, Fatemehsadat Saleh

Despite the recent advances in multiple object tracking (MOT), achieved by joint detection and tracking, dealing with long occlusions remains a challenge. This is due to the fact that such techniques tend to ignore the long-term motion information. In this ...

IEEE COMPUTER SOC2021

Generating Smooth Pose Sequences for Diverse Human Motion Prediction

Mathieu Salzmann, Wei Mao

Recent progress in stochastic motion prediction, i.e., predicting multiple possible future human motions given a single past pose sequence, has led to producing truly diverse future motions and even providing control over the motion of some body parts. How ...

IEEE2021

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

Mathieu Salzmann, Yinlin Hu, Zhigang Li

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

Landmark Regularization: Ranking Guided Super-Net Training in Neural Architecture Search

Mathieu Salzmann, Kaicheng Yu

Weight sharing has become a de facto standard in neural architecture search because it enables the search to be done on commodity hardware. However, recent works have empirically shown a ranking disorder between the performance of stand-alone architectures ...

IEEE COMPUTER SOC2021

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

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

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

Page 4 of 9