Publication

MixCycle: Mixup Assisted Semi-Supervised 3D Single Object Tracking with Cycle Consistency

Publications associées (28)

Highly Hygroscopic Aerosols Facilitate Summer and Early‐Autumn Cloud Formation at Extremely Low Concentrations Over the Central Arctic Ocean

Arctic clouds are sensitive to atmospheric particles since these are sometimes in such low concentrations that clouds cannot always form under supersaturated water vapor conditions. This is especially true in the late summer, when aerosol concentrations ar ...

2024

Modular Multiplatform Compatible Air Measurement System (MoMuCAMS): a new modular platform for boundary layer aerosol and trace gas vertical measurements in extreme environments

Roman Pohorsky, Julia Schmale, Andrea Baccarini

The Modular Multiplatform Compatible Air Measurement System (MoMuCAMS) is a newly developed in situ aerosol and trace gas measurement platform for lower-atmospheric vertical profiling. MoMuCAMS has been primarily designed to be attached to a Helikite, a ru ...

2024

Simulations of primary and secondary ice production during an Arctic mixed-phase cloud case from the Ny-Ålesund Aerosol Cloud Experiment (NASCENT) campaign

Georgia Sotiropoulou, Paraskevi Georgakaki

The representation of Arctic clouds and their phase distributions, i.e., the amount of ice and supercooled water, influences predictions of future Arctic warming. Therefore, it is essential that cloud phase is correctly captured by models in order to accur ...

Copernicus GmbH2024

Match Normalization: Learning-Based Point Cloud Registration for 6D Object Pose Estimation in the Real World

Mathieu Salzmann, Zheng Dang

In this work, we tackle the task of estimating the 6D pose of an object from point cloud data. While recent learning-based approaches have shown remarkable success on synthetic datasets, we have observed them to fail in the presence of real-world data. We ...

Ieee Computer Soc2024

Characterization of Transverse Instabilities Driven by Electron Cloud

Sofia Carolina Johannesson

Electron cloud continues to be one of the main limiting factors of the Large Hadron Collider (LHC), the biggest accelerator at CERN. These clouds form in the beam chamber when positively charged particles are passing through and cause unwanted effects in b ...

EPFL2024

The Skin of Venice: Automatic Facade Extraction from Point Clouds

Isabella Di Lenardo, Paul Robert Guhennec

We propose a method to extract orthogonal views of facades from photogrammetric models of cities. This method was applied to extract all facades of the city of Venice. The result images open up new areas of research in architectural history. ...

Zentrum für Informationsmodellierung - Austrian Centre for Digital Humanities, University of Graz2023

Aerosol and dynamical contributions to cloud droplet formation in Arctic low-level clouds

Guangyu Li, Athanasios Nenes, Jörg Wieder, Paraskevi Georgakaki, Ghislain Gilles Jean-Michel Motos

The Arctic is one of the most rapidly warming regions of the globe. Low-level clouds and fog modify the energy transfer from and to space and play a key role in the observed strong Arctic surface warming, a phenomenon commonly termed "Arctic amplification" ...

Copernicus GmbH2023

Role of K-feldspar and quartz in global ice nucleation by mineral dust in mixed-phase clouds

Athanasios Nenes

Ice-nucleating particles (INPs) enable ice formation, profoundly affecting the microphysical and radiative properties, lifetimes, and precipitation rates of clouds. Mineral dust emitted from arid regions, particularly potassium-containing feldspar (K-felds ...

Copernicus GmbH2023

The role of secondary ice production in mixed-phase clouds

Paraskevi Georgakaki

Clouds are omnipresent in the Earth's atmosphere. Their phase composition significantly modulates their interaction with solar and terrestrial radiation, as well as precipitation formation. Particularly for clouds containing both phases, known as mixed-pha ...

EPFL2023

Multi-Source Fusion for Voxel-Based 7-DoF Grasping Pose Estimation

Fei Wang, Zheng Dang

In this work, we tackle the problem of 7-DoF grasping pose estimation(6-DoF with the opening width of parallel-jaw gripper) from point cloud data, which is a fundamental task in robotic manipulation. Most existing methods adopt 3D voxel CNNs as the backbon ...

IEEE2023

Brown carbon absorption in the Mediterranean basin from local and long-range transported biomass burning air masses

Athanasios Nenes

This study examines the impact of light absorption from biomass burning (BB) brown carbon (BrC) in the Mediterranean basin from local and distant fire incidents during a typical fire season in August 2019 and under severe fire activity in August 2021. The ...

Elsevier2023

A Case Study on Drivers of the Isotopic Composition of Water Vapor at the Coast of East Antarctica

Michael Lehning, Armin Sigmund, Riqo Chaar

Stable water isotopes (SWIs) contain valuable information on the past climate and phase changes in the hydrologic cycle. Recently, vapor measurements in the polar regions have provided new insights into the effects of snow-related and atmospheric processes ...

AMER GEOPHYSICAL UNION2023

Reducing Aerosol Forcing Uncertainty by Combining Models With Satellite and Within-The-Atmosphere Observations: A Three-Way Street

Athanasios Nenes

Aerosol forcing uncertainty represents the largest climate forcing uncertainty overall. Its magnitude has remained virtually undiminished over the past 20 years despite considerable advances in understanding most of the key contributing elements. Recent wo ...

AMER GEOPHYSICAL UNION2023

Calibration Transfer Methodology for Cloud Radars Based on Ice Cloud Observations

Alexis Berne, Anne-Claire Marie Billault--Roux

This article presents a calibration transfer methodology that can be used between radars of the same or dif-ferent frequency bands. This method enables the absolute calibration of a cloud radar by transferring it from another collocated instrument with kno ...

AMER METEOROLOGICAL SOC2023

Robust Training and Verification of Deep Neural Networks

Fabian Ricardo Latorre Gomez

According to the proposed Artificial Intelligence Act by the European Comission (expected to pass at the end of 2023), the class of High-Risk AI Systems (Title III) comprises several important applications of Deep Learning like autonomous driving vehicles ...

EPFL2023

Encoder-Decoder Models for Human Segmentation and Motion Analysis

Isinsu Katircioglu

Detecting people from 2D images and analyzing their motion in 3D have been long standing computer vision problems central to numerous applications such as autonomous driving and athletic training. Recently, with the availability of large amounts of trainin ...

EPFL2022

Reshaping Perception for Autonomous Driving with Semantic Keypoints

Lorenzo Bertoni

The field of artificial intelligence is set to fuel the future of mobility by driving forward the transition from advanced driver-assist systems to fully autonomous vehicles (AV). Yet the current technology, backed by cutting-edge deep learning techniques, ...

EPFL2022

An analytical approximation of urban heat and dry islands and their impact on convection triggering

Kai Wang, Gabriele Manoli

It is well known that cities increase air and surface temperatures compared to their rural sur-roundings, the so-called urban heat island (UHI) effect. However, the associated changes in at-mospheric humidity (also known as urban dry island, UDI) and conve ...

ELSEVIER2022

Incompleteness of graph neural networks for points clouds in three dimensions

Michele Ceriotti, Sergey Pozdnyakov

Graph neural networks (GNN) are very popular methods in machine learning and have been applied very successfully to the prediction of the properties of molecules and materials. First-order GNNs are well known to be incomplete, i.e. there exist graphs that ...

IOP Publishing Ltd2022

Weakly Supervised Volumetric Image Segmentation with Deformed Templates

Pascal Fua, Pamuditha Udaranga Wickramasinghe, Patrick Moller Jensen, Mian Akbar Shah

There are many approaches to weakly-supervised training of networks to segment 2D images. By contrast, existing approaches to segmenting volumetric images rely on full-supervision of a subset of 2D slices of the 3D volume. We propose an approach to volume ...

Springer2022