no code implementations • 22 May 2024 • Luca Savant Aira, Antonio Montanaro, Emanuele Aiello, Diego Valsesia, Enrico Magli
Generating videos with realistic and physically plausible motion is one of the main recent challenges in computer vision.
no code implementations • 27 Mar 2024 • Luca Savant, Diego Valsesia, Enrico Magli
We present Stochastic Gaussian Splatting (SGS): the first framework for uncertainty estimation using Gaussian Splatting (GS).
no code implementations • 26 Mar 2024 • Diego Valsesia, Tiziano Bianchi, Enrico Magli
Deep learning methods have traditionally been difficult to apply to compression of hyperspectral images onboard of spacecrafts, due to the large computational complexity needed to achieve adequate representational power, as well as the lack of suitable datasets for training and testing.
no code implementations • 30 Jan 2024 • Luca Savant Aira, Diego Valsesia, Andrea Bordone Molini, Giulia Fracastoro, Enrico Magli, Andrea Mirabile
Multi-image super-resolution (MISR) allows to increase the spatial resolution of a low-resolution (LR) acquisition by combining multiple images carrying complementary information in the form of sub-pixel offsets in the scene sampling, and can be significantly more effective than its single-image counterpart.
1 code implementation • 19 Jan 2023 • Emanuele Aiello, Diego Valsesia, Enrico Magli
Our findings show that the proposed method is able to produce high-quality samples in a fraction of the time required by widely-used diffusion models, and outperforms state-of-the-art techniques for accelerated sampling.
1 code implementation • 21 Sep 2022 • Antonio Montanaro, Diego Valsesia, Enrico Magli
In this paper, we propose to embed the features of a point cloud classifier into the hyperbolic space and explicitly regularize the space to account for the part-whole hierarchy.
Ranked #9 on 3D Point Cloud Classification on ModelNet40
1 code implementation • 20 Sep 2022 • Emanuele Aiello, Diego Valsesia, Enrico Magli
In this paper we explore the recent topic of point cloud completion, guided by an auxiliary image.
Ranked #1 on Point Cloud Completion on ShapeNet-ViPC
no code implementations • 1 Jul 2022 • Antonio Montanaro, Diego Valsesia, Enrico Magli
Inverse problems consist in reconstructing signals from incomplete sets of measurements and their performance is highly dependent on the quality of the prior knowledge encoded via regularization.
no code implementations • 6 Apr 2022 • Diego Valsesia, Enrico Magli
Multi-image super-resolution from multi-temporal satellite acquisitions of a scene has recently enjoyed great success thanks to new deep learning models.
no code implementations • 20 Aug 2021 • Antonio Montanaro, Diego Valsesia, Giulia Fracastoro, Enrico Magli
Semi-supervised learning techniques are gaining popularity due to their capability of building models that are effective, even when scarce amounts of labeled data are available.
1 code implementation • 26 May 2021 • Diego Valsesia, Enrico Magli
However, existing models have neglected the issue of temporal permutation, whereby the temporal ordering of the input images does not carry any relevant information for the super-resolution task and causes such models to be inefficient with the, often scarce, ground truth data that available for training.
Ranked #2 on Multi-Frame Super-Resolution on PROBA-V
1 code implementation • CVPR 2021 • Antonio Alliegro, Diego Valsesia, Giulia Fracastoro, Enrico Magli, Tatiana Tommasi
The combined embedding inherits category-agnostic properties from the chosen pretext tasks.
no code implementations • 29 Mar 2021 • Diego Valsesia, Giulia Fracastoro, Enrico Magli
In this paper, we investigate the recently proposed randomly wired architectures in the context of graph neural networks.
Ranked #20 on Graph Property Prediction on ogbg-molpcba
no code implementations • ICLR Workshop GTRL 2021 • Diego Valsesia, Giulia Fracastoro, Enrico Magli
In this paper, we investigate the recently proposed randomly wired architectures in the context of graph neural networks.
no code implementations • 10 Dec 2020 • Giulia Fracastoro, Enrico Magli, Giovanni Poggi, Giuseppe Scarpa, Diego Valsesia, Luisa Verdoliva
Synthetic aperture radar (SAR) images are affected by a spatially-correlated and signal-dependent noise called speckle, which is very severe and may hinder image exploitation.
no code implementations • 29 Oct 2020 • Arslan Ali, Andrea Migliorati, Tiziano Bianchi, Enrico Magli
Deep learning has shown outstanding performance in several applications including image classification.
no code implementations • ECCV 2020 • Arslan Ali, Matteo Testa, Tiziano Bianchi, Enrico Magli
We present BioMetricNet: a novel framework for deep unconstrained face verification which learns a regularized metric to compare facial features.
1 code implementation • ECCV 2020 • Francesca Pistilli, Giulia Fracastoro, Diego Valsesia, Enrico Magli
Point clouds are an increasingly relevant data type but they are often corrupted by noise.
1 code implementation • 4 Jul 2020 • Andrea Bordone Molini, Diego Valsesia, Giulia Fracastoro, Enrico Magli
Information extraction from synthetic aperture radar (SAR) images is heavily impaired by speckle noise, hence despeckling is a crucial preliminary step in scene analysis algorithms.
no code implementations • 15 Jan 2020 • Andrea Bordone Molini, Diego Valsesia, Giulia Fracastoro, Enrico Magli
The proposed method is trained employing only noisy images and can therefore learn features of real SAR images rather than synthetic data.
no code implementations • 15 Jan 2020 • Andrea Bordone Molini, Diego Valsesia, Giulia Fracastoro, Enrico Magli
Deep learning methods for super-resolution of a remote sensing scene from multiple unregistered low-resolution images have recently gained attention thanks to a challenge proposed by the European Space Agency.
no code implementations • 20 Nov 2019 • Matteo Testa, Arslan Ali, Tiziano Bianchi, Enrico Magli
Differently from other methods, RegNet learns a mapping of the input biometric traits onto a target distribution in a well-behaved space in which users can be separated by means of simple and tunable boundaries.
no code implementations • 2 Sep 2019 • Diego Valsesia, Sophie Marie Fosson, Chiara Ravazzi, Tiziano Bianchi, Enrico Magli
Embeddings provide compact representations of signals in order to perform efficient inference in a wide variety of tasks.
1 code implementation • 19 Jul 2019 • Diego Valsesia, Giulia Fracastoro, Enrico Magli
The graph convolution operation generalizes the classic convolution to arbitrary graphs.
Ranked #3 on Grayscale Image Denoising on Set12 sigma25
1 code implementation • 15 Jul 2019 • Andrea Bordone Molini, Diego Valsesia, Giulia Fracastoro, Enrico Magli
This novel framework integrates the spatial registration task directly inside the CNN, and allows to exploit the representation learning capabilities of the network to enhance registration accuracy.
Ranked #8 on Multi-Frame Super-Resolution on PROBA-V
1 code implementation • 5 Jul 2019 • Diego Valsesia, Enrico Magli
Compression of hyperspectral images onboard of spacecrafts is a tradeoff between the limited computational resources and the ever-growing spatial and spectral resolution of the optical instruments.
1 code implementation • 29 May 2019 • Diego Valsesia, Giulia Fracastoro, Enrico Magli
The graph-convolutional layers dynamically construct neighborhoods in the feature space to detect latent correlations in the feature maps produced by the hidden layers.
1 code implementation • ICLR 2019 • Diego Valsesia, Giulia Fracastoro, Enrico Magli
We also study the problem of defining an upsampling layer in the graph-convolutional generator, such that it learns to exploit a self-similarity prior on the data distribution to sample more effectively.
no code implementations • 7 Jul 2017 • Attilio Fiandrotti, Sophie M. Fosson, Chiara Ravazzi, Enrico Magli
Finally, we practically demonstrate our algorithms in a typical application of circulant matrices: deblurring a sparse astronomical image in the compressed domain.
no code implementations • 30 Jan 2017 • Diego Valsesia, Enrico Magli
We use some of the largest order statistics of the random projections of a reference signal to construct a binary embedding that is adapted to signals correlated with such signal.
no code implementations • 12 Oct 2016 • Pedro Porto Buarque de Gusmão, Gianluca Francini, Skjalg Lepsøy, Enrico Magli
Training deep Convolutional Neural Networks (CNN) is a time consuming task that may take weeks to complete.
no code implementations • 7 Mar 2014 • Simeon Kamdem Kuiteing, Giulio Coluccia, Alessandro Barducci, Mauro Barni, Enrico Magli
Compressed Sensing (CS) is suitable for remote acquisition of hyperspectral images for earth observation, since it could exploit the strong spatial and spectral correlations, llowing to simplify the architecture of the onboard sensors.
no code implementations • 4 Nov 2013 • Tomas Björklund, Enrico Magli
Simulations of our architecture show that the image quality is comparable to that of a classic Compressive Imaging camera, whereas the proposed architecture avoids long acquisition times due to sequential sensing.
no code implementations • 8 Oct 2013 • Giulio Coluccia, Diego Valsesia, Enrico Magli
On one hand, it allows to enforce a set of constraints to drive the reconstruction algorithm towards a smooth solution, imposing the similarity of block borders.
no code implementations • 4 Oct 2013 • Giulio Coluccia, Enrico Magli
The use of CS in such setting raises the problem of reconstructing a very high number of samples, as are contained in an image, from their linear projections.
no code implementations • 4 Oct 2013 • Diego Valsesia, Enrico Magli
The proposed method successfully provides resolution and quality scalability with modest complexity and it provides gains in the quality of the reconstructed images with respect to separate encoding of the quality layers.