Publications associées à On the impact of learning-based image compression on computer vision tasks

Publications associées (25)

Investigating OCR-Sensitive Neurons to Improve Entity Recognition in Historical Documents

This paper investigates the presence of OCR-sensitive neurons within the Transformer architecture and their influence on named entity recognition (NER) performance on historical documents. By analysing neuron activation patterns in response to clean and no ...

Springer Science and Business Media Deutschland GmbH2025

Augmented Carpentry: Integration of Manual Fabrication into the Digital Value Chain

Andrea Settimi

Full robotic automation in construction and digital fabrication has gained significant momentum in the last 20 years. Yet, although effective, these technologies demand considerable capital investment, highly qualified operators, and are often tailored to ...

EPFL2025

Automatic Parameterization for Aerodynamic Shape Optimization via Deep Geometric Learning

Pascal Fua, Zhen Wei

We propose two deep learning models that fully automate shape parameterization for aerodynamic shape optimization. Both models are optimized to parameterize via deep geometric learning to embed human prior knowledge into learned geometric patterns, elimina ...

2023

Estimating cross-field particle transport at the outer midplane of TCV by tracking filaments with machine learning

Christian Gabriel Theiler, Davide Galassi, Claudia Colandrea, Curdin Tobias Wüthrich, Nicola Offeddu

Cross-field transport of particles in the boundary region of magnetically confined fusion plasmas is dominated by turbulence. Blobs, intermittent turbulent structures with large amplitude and a filamentary shape appearing in the scrape-off layer (SOL), are ...

IOP Publishing Ltd2023

Deep transfer learning of additive manufacturing mechanisms across materials in metal-based laser powder bed fusion process

Roland Logé, Rita Drissi Daoudi

The defective regimes in metal-based Laser Powder Bed Fusion (LPBF) processes can be minimized by deploying in-situ monitoring strategies comprising Machine learning (ML) algorithms and sensing techniques. So far, algorithms trained for monitoring a partic ...

ELSEVIER SCIENCE SA2022

Perceiving Humans: from Monocular 3D Localization to Social Distancing

Alexandre Massoud Alahi, Sven Kreiss, Lorenzo Bertoni

Perceiving humans in the context of Intelligent Transportation Systems (ITS) often relies on multiple cameras or expensive LiDAR sensors. In this work, we present a new cost-effective vision-based method that perceives humans' locations in 3D and their bod ...

IEEE Institute of Electrical and Electronics Engineers2022

Brain signals of a Surprise-Actor-Critic model: Evidence for multiple learning modules in human decision making

Wulfram Gerstner, Johanni Michael Brea, Kerstin Preuschoff, Alireza Modirshanechi, Marco Philipp Lehmann, Vasiliki Liakoni

Learning how to reach a reward over long series of actions is a remarkable capability of humans, and potentially guided by multiple parallel learning modules. Current brain imaging of learning modules is limited by (i) simple experimental paradigms, (ii) e ...

ACADEMIC PRESS INC ELSEVIER SCIENCE2022

Active Learning for Imbalanced Civil Infrastructure Data

Diego Matteo Antognini, Adelmo Cristiano Innocenza Malossi

Aging civil infrastructures are closely monitored by engineers for damage and critical defects. As the manual inspection of such large structures is costly and time-consuming, we are working towards fully automating the visual inspections to support the pr ...

2022

JPEG XS-A New Standard for Visually Lossless Low-Latency Lightweight Image Coding

Touradj Ebrahimi

Joint Photographic Experts Group (JPEG) XS is a new International Standard from the JPEG Committee (formally known as ISO/International Electrotechnical Commission (IEC) JTC1/SC29/WG1). It defines an interoperable, visually lossless low-latency lightweight ...

IEEE-INST ELECTRICAL ELECTRONICS ENGINEERS INC2021

Robustness via Cross-Domain Ensembles

Amir Roshan Zamir, Shuqing Teresa Yeo, Oguzhan Fatih Kar

We present a method for making neural network predictions robust to shifts from the training data distribution. The proposed method is based on making predictions via a diverse set of cues (called 'middle domains') and ensembling them into one strong predi ...

IEEE2021

Training Channel Selection for Learning-based 1-bit Precoding in Massive MU-MIMO

Andreas Peter Burg, Alexios Konstantinos Balatsoukas Stimming, Sitian Li

Learning-based algorithms have gained great popularity in communications since they often outperform even carefully engineered solutions by learning from training samples. In this paper, we show that the selection of appropriate training examples can be im ...

IEEE2020

On Symmetry and Initialization for Neural Networks

Ido Nachum

This work provides an additional step in the theoretical understanding of neural networks. We consider neural networks with one hidden layer and show that when learning symmetric functions, one can choose initial conditions so that standard SGD training ef ...

Springer Nature2020

Generating Master Faces for Use in PerformingWolf Attacks on Face Recognition Systems

Sébastien Marcel

Due to its convenience, biometric authentication, especial face authentication, has become increasingly mainstream and thus is now a prime target for attackers. Presentation attacks and face morphing are typical types of attack. Previous research has shown ...

IEEE2020

An HMM Approach with Inherent Model Selection for Sign Language and Gesture Recognition

Mathew Magimai Doss, Sandrine Tornay

HMMs have been the one of the first models to be applied for sign recognition and have become the baseline models due to their success in modeling sequential and multivariate data. Despite the extensive use of HMMs for sign recognition, determining the HMM ...

EUROPEAN LANGUAGE RESOURCES ASSOC-ELRA2020

Detecting the Unexpected via Image Resynthesis

Pascal Fua, Mathieu Salzmann, Krzysztof Maciej Lis, Krishna Kanth Nakka

Classical semantic segmentation methods, including the recent deep learning ones, assume that all classes observed at test time have been seen during training. In this paper, we tackle the more realistic scenario where unexpected objects of unknown classes ...

IEEE2019

A Performance Evaluation of Local Features for Image-Based 3D Reconstruction

Pascal Fua, Bin Fan

This paper performs a comprehensive and comparative evaluation of the state of the art local features for the task of image based 3D reconstruction. The evaluated local features cover the recently developed ones by using powerful machine learning technique ...

2019

A new objective metric to predict image quality using deep neural networks

Touradj Ebrahimi, Pinar Akyazi

Quality assessment of images is of key importance for media applications. In this paper we present a new objective metric to predict the quality of images using deep neural networks. The network makes use of both the color information as well as frequency ...

2018

Full-Pulse Tomographic Reconstruction with Deep Neural Networks

Olivier Sauter, Martinus Adela Maria Gijs, Jonathan Graves, Ambrogio Fasoli, Stefano Coda, Basil Duval, Henri Weisen, Richard Pitts, Yves Martin, Javier García Hernández, Duccio Testa, Miguel Fernández Ruiz, Nicola Vianello, Robin Humphry-Baker, Sun Hee Kim, Federico Nespoli, Patrick Blanchard, Alessandro Pau, David Pfefferlé, Davide Galassi, Jonathan Marc Philippe Faustin, Cristian Sommariva, Hamish William Patten, Samuel Lanthaler, Jan Horacek, Yann Camenen, Bruno Emanuel Ferreira De Sousa Correia, José Pedro Rebelo Ferreira Marques, Mikhail Maslov, Marco Wischmeier, Dalziel Joseph Wilson, Liang Yao, Daniel Scott Alessi, Arnout Lodewijk M Beckers, Ana Francisca Leal Silva Soares, Jonnathan Cesar Hidalgo Acosta, Antonio José Pereira de Figueiredo, Partha Dutta, Pierre-Thomas Paul Brun, Pedro Camilo de Oliveira e Silva, Alberto Hernando de Castro, Julio Rodriguez, Vlad Trifa, Li Shuai, Rebecca Hill

Plasma tomography consists of reconstructing a two-dimensional radiation profile of a poloidal cross section of a fusion device based on line-integrated measurements along several lines of sight. The reconstruction process is computationally intensive, and ...

2018

Algorithmic Composition of Melodies with Deep Recurrent Neural Networks

Wulfram Gerstner, Johanni Michael Brea, Samuel Pavio Muscinelli, Alexander Kevin Seeholzer, Florian François Colombo

A big challenge in algorithmic composition is to devise a model that is both easily trainable and able to reproduce the long-range temporal dependencies typical of music. Here we investigate how artificial neural networks can be trained on a large corpus o ...

2016

Automated Reconstruction of Evolving Curvilinear Tree Structures

Przemyslaw Rafal Glowacki

Curvilinear networks are prevalent in nature and span many different scales, ranging from micron-scale neural structures in the brain to petameter-scale dark-matter arbors binding massive galaxy clusters. Reliably reconstructing them in an automated fashio ...

EPFL2016