|
Nando Metzger I am a PhD student in the Photogrammetry and Remote Sensing Lab, ETH Zürich My supervisors are Prof. Konrad Schindler and Prof. Devis Tuia. I work on computer vision problems with various applications such as monocular depth estimation and remote sensing. I am collaborating with the ICRC to map vulnerable populations in developing countries. Previously, interned at Meta's Reality Labs Research, with which I am still collaborating with. I obtained my bachelor and master's degree in Geomatics Engineering from ETH Zürich. During my master's specialized in deep learning, computer vision and remote sensing. |
|
|
|
|
|
News
|
ResearchI'm interested in computer vision, deep learning, and their applications to remote sensing. Most of my work is related to super-resolution, depth estimation or both at the same time. Some papers are highlighted. * indicates equal contribution. |
|
Marigold: Repurposing Diffusion-Based Image Generators for Monocular Depth Estimation
Bingxin Ke, Anton Obukhov, Shengyu Huang, Nando Metzger, Rodrigo Caye Daudt, Konrad Schindler, CVPR, 2024 (Oral, Best Paper Candidate) project page / arXiv / code / colab / demo Marigold is an affine-invariant monocular depth estimation method based on Stable Diffusion, leveraging its rich prior knowledge for better generalization and achieving state-of-the-art performance with significant improvements, even with synthetic training data. |
|
Thera: Neural Fields with Thermal Activations for Arbitrary-Scale Super-Resolution
Alexander Becker*, Rodrigo Caye Daudt*, Nando Metzger, Jan Dirk Wegner, Konrad Schindler, arXiv, 2023 arXiv / code We present a novel approach for arbitrary-scale single image super-resolution (ASSR) that uses neural fields with an adaptive Gaussian point spread function (PSF) to prevent aliasing and achieve superior results, offering more parameter efficiency and setting a new state of the art while maintaining computational efficiency. |
|
POPCORN: High-resolution Population Maps Derived from Sentinel-1 and Sentinel-2
Nando Metzger, Rodrigo Caye Daudt, Devis Tuia Konrad Schindler, arXiv, 2023 project page / code / arXiv / demo / data POPCORN is a lightweight population mapping method using free satellite images and minimal data, surpassing existing accuracy and providing interpretable maps for mapping populations in data-scarce regions. |
|
DADA: Guided Depth Super-Resolution by Deep Anisotropic Diffusion
Nando Metzger*, Rodrigo Caye Daudt*, Konrad Schindler, CVPR, 2023 arXiv / paper / project page / video / poster We propose DADA, a novel approach to depth image super-resolution by combining guided anisotropic diffusion with a deep convolutional network, enhancing both edge detail and contextual reasoning. This method achieves unprecedented results in three benchmarks, especially at larger scales like x32 |
|
POMELO: Fine-grained Population Mapping from Coarse Census Counts and Open Geodata
Nando Metzger, John E Vargas-Muñoz, Rodrigo Caye Daudt, Benjamin Kellenberger, Thao Ton-That Whelan, Muhammad Imran, Ferda Ofli, Konrad Schindler, Devis Tuia Nature - Scientific Reports, 2022 project page / video / arXiv POMELO is a deep learning model that creates fine-grained population maps using coarse census counts and open geodata, achieving high accuracy in sub-Saharan Africa and effectively estimating population numbers even without any census data. |
|
Urban Change Forecasting from Satellite Images
Nando Metzger, Mehmet Ozgur Turkoglu Rodrigo Caye Daudt, Jan Dirk Wegner, Konrad Schindler, PFG, 2022, (Karl-Kraus Best Paper Award) paper We propose a method for forecasting the emergence and timing of new buildings using a deep neural network with a custom pretraining procedure, validated on the SpaceNet7 dataset. |
|
Crop Classification under Varying Cloud Cover with Neural Ordinary Differential Equations
Nando Metzger*, Mehmet Ozgur Turkoglu*, Stefano D'Aronco, Jan Dirk Wegner, Konrad Schindler, IEEE, TGRS, 2021, paper We propose using neural ordinary differential equations (NODEs) combined with RNNs to improve crop classification from irregularly spaced satellite images, showing enhanced accuracy over common methods, especially with few observations, and better early-season forecasting due to the continuous representation of latent dynamics. |
|
DSM Refinement with Deep Encoder-Decoder Networks
Nando Metzger, Corinne Stucker, Konrad Schindler, arXiv, 2020, paper This work presents a method for automatically refining 3D city models generated from aerial images by using a neural network trained with reference data and a loss function to improve DSMs, effectively preserving geometric structures while removing noise and artifacts. |
|
Thank you for the template Jon Barron. |