no code implementations • 18 Mar 2024 • Lisa Weijler, Muhammad Jehanzeb Mirza, Leon Sick, Can Ekkazan, Pedro Hermosilla
Given access to paired image-pointcloud (2D-3D) data, we first optimize a 3D segmentation backbone for the main task of semantic segmentation using the pointclouds and the task of 2D $\to$ 3D KD by using an off-the-shelf 2D pre-trained foundation model.
no code implementations • 23 Feb 2024 • Leon Sick, Dominik Engel, Pedro Hermosilla, Timo Ropinski
Masked autoencoders (MAEs) have established themselves as a powerful method for unsupervised pre-training for computer vision tasks.
no code implementations • 19 Feb 2024 • Sebastian Koch, Narunas Vaskevicius, Mirco Colosi, Pedro Hermosilla, Timo Ropinski
We co-embed the features from a 3D scene graph prediction backbone with the feature space of powerful open world 2D vision language foundation models.
1 code implementation • 16 Jan 2024 • Philipp Erler, Lizeth Fuentes, Pedro Hermosilla, Paul Guerrero, Renato Pajarola, Michael Wimmer
3D surface reconstruction from point clouds is a key step in areas such as content creation, archaeology, digital cultural heritage, and engineering.
1 code implementation • 6 Nov 2023 • Lisa Weijler, Florian Kowarsch, Michael Reiter, Pedro Hermosilla, Margarita Maurer-Granofszky, Michael Dworzak
This limitation becomes particularly relevant in scenarios where the attributes captured during data acquisition vary across different samples.
no code implementations • 25 Oct 2023 • Sebastian Koch, Pedro Hermosilla, Narunas Vaskevicius, Mirco Colosi, Timo Ropinski
While it is widely accepted that pre-training is an effective approach to improve model performance in low data regimes, in this paper, we find that existing pre-training methods are ill-suited for 3D scene graphs.
no code implementations • 3 Oct 2023 • Pedro Hermosilla
Additionally, we show that a neural network architecture using simple convolutions based on such embeddings is able to achieve state-of-the-art results on several tasks, outperforming recent and more complex operations.
no code implementations • 27 Sep 2023 • Sebastian Koch, Pedro Hermosilla, Narunas Vaskevicius, Mirco Colosi, Timo Ropinski
In the field of 3D scene understanding, 3D scene graphs have emerged as a new scene representation that combines geometric and semantic information about objects and their relationships.
no code implementations • 21 Sep 2023 • Leon Sick, Dominik Engel, Pedro Hermosilla, Timo Ropinski
We achieve this by (1) learning depth-feature correlation by spatially correlate the feature maps with the depth maps to induce knowledge about the structure of the scene and (2) implementing farthest-point sampling to more effectively select relevant features by utilizing 3D sampling techniques on depth information of the scene.
no code implementations • 27 Apr 2023 • Sebastian Hartwig, Christian van Onzenoodt, Dominik Engel, Pedro Hermosilla, Timo Ropinski
Finally, we compare our approach against existing state-of-the-art clustering techniques and can show, that ClusterNet is able to generalize to unseen and out of scope data.
1 code implementation • 11 Oct 2022 • Michael Schelling, Pedro Hermosilla, Timo Ropinski
Indirect Time-of-Flight (iToF) cameras are a widespread type of 3D sensor, which perform multiple captures to obtain depth values of the captured scene.
no code implementations • 31 May 2022 • Pedro Hermosilla, Timo Ropinski
Learning from 3D protein structures has gained wide interest in protein modeling and structural bioinformatics.
1 code implementation • CVPR 2022 • Hannah Kniesel, Timo Ropinski, Tim Bergner, Kavitha Shaga Devan, Clarissa Read, Paul Walther, Tobias Ritschel, Pedro Hermosilla
Scanning Transmission Electron Microscopes (STEMs) acquire 2D images of a 3D sample on the scale of individual cell components.
1 code implementation • ICLR 2022 • Adam Celarek, Pedro Hermosilla, Bernhard Kerbl, Timo Ropinski, Michael Wimmer
This paper proposes a novel method for deep learning based on the analytical convolution of multidimensional Gaussian mixtures.
no code implementations • 7 Dec 2021 • Pedro Hermosilla, Michael Schelling, Tobias Ritschel, Timo Ropinski
Appropriate weight initialization has been of key importance to successfully train neural networks.
1 code implementation • CVPR 2022 • Michael Schelling, Pedro Hermosilla, Timo Ropinski
Time-of-Flight (ToF) cameras are subject to high levels of noise and distortions due to Multi-Path-Interference (MPI).
1 code implementation • 23 Jul 2021 • Patrik Puchert, Pedro Hermosilla, Tobias Ritschel, Timo Ropinski
Density estimation plays a crucial role in many data analysis tasks, as it infers a continuous probability density function (PDF) from discrete samples.
1 code implementation • ICLR 2021 • Pedro Hermosilla, Marco Schäfer, Matěj Lang, Gloria Fackelmann, Pere Pau Vázquez, Barbora Kozlíková, Michael Krone, Tobias Ritschel, Timo Ropinski
Proteins perform a large variety of functions in living organisms, thus playing a key role in biology.
1 code implementation • 10 Mar 2020 • Michael Schelling, Pedro Hermosilla, Pere-Pau Vazquez, Timo Ropinski
Optimal viewpoint prediction is an essential task in many computer graphics applications.
1 code implementation • ICCV 2019 • Pedro Hermosilla, Tobias Ritschel, Timo Ropinski
We show that denoising of 3D point clouds can be learned unsupervised, directly from noisy 3D point cloud data only.
1 code implementation • 12 Nov 2018 • Pedro Hermosilla, Sebastian Maisch, Tobias Ritschel, Timo Ropinski
Thus, we suggest a two-stage operator comprising of a 3D network that first transforms the point cloud into a latent representation, which is later on projected to the 2D output image using a dedicated 3D-2D network in a second step.
1 code implementation • 5 Jun 2018 • Pedro Hermosilla, Tobias Ritschel, Pere-Pau Vázquez, Àlvar Vinacua, Timo Ropinski
We propose an efficient and effective method to learn convolutions for non-uniformly sampled point clouds, as they are obtained with modern acquisition techniques.