Search Results for author: Pietro Morerio

Found 39 papers, 20 papers with code

Leveraging Acoustic Images for Effective Self-Supervised Audio Representation Learning

1 code implementation ECCV 2020 Valentina Sanguineti, Pietro Morerio, Niccolò Pozzetti, Danilo Greco, Marco Cristani, Vittorio Murino

However, since 2D planar arrays are cumbersome and not as widespread as ordinary microphones, we propose that the richer information content of acoustic images can be distilled, through a self-supervised learning scheme, into more powerful audio and visual feature representations.

Cross-Modal Retrieval Representation Learning +3

DiffAssemble: A Unified Graph-Diffusion Model for 2D and 3D Reassembly

1 code implementation29 Feb 2024 Gianluca Scarpellini, Stefano Fiorini, Francesco Giuliari, Pietro Morerio, Alessio Del Bue

Reassembly tasks play a fundamental role in many fields and multiple approaches exist to solve specific reassembly problems.

Denoising

Learnable Data Augmentation for One-Shot Unsupervised Domain Adaptation

1 code implementation3 Oct 2023 Julio Ivan Davila Carrazco, Pietro Morerio, Alessio Del Bue, Vittorio Murino

This paper presents a classification framework based on learnable data augmentation to tackle the One-Shot Unsupervised Domain Adaptation (OS-UDA) problem.

Data Augmentation One-shot Unsupervised Domain Adaptation +2

Leveraging Next-Active Objects for Context-Aware Anticipation in Egocentric Videos

no code implementations16 Aug 2023 Sanket Thakur, Cigdem Beyan, Pietro Morerio, Vittorio Murino, Alessio Del Bue

Compared to existing video modeling architectures for action anticipation, NAOGAT captures the relationship between objects and the global scene context in order to predict detections for the next active object and anticipate relevant future actions given these detections, leveraging the objects' dynamics to improve accuracy.

Action Anticipation Active Object Localization +3

Person Re-Identification without Identification via Event Anonymization

1 code implementation ICCV 2023 SHAFIQ AHMAD, Pietro Morerio, Alessio Del Bue

In this work, we also bring to the community the first ever event-based person ReId dataset gathered to evaluate the performance of our approach.

Event-based vision Image Reconstruction +2

Guided Attention for Next Active Object @ EGO4D STA Challenge

1 code implementation25 May 2023 Sanket Thakur, Cigdem Beyan, Pietro Morerio, Vittorio Murino, Alessio Del Bue

In this technical report, we describe the Guided-Attention mechanism based solution for the short-term anticipation (STA) challenge for the EGO4D challenge.

Object Short-term Object Interaction Anticipation

Enhancing Next Active Object-based Egocentric Action Anticipation with Guided Attention

1 code implementation22 May 2023 Sanket Thakur, Cigdem Beyan, Pietro Morerio, Vittorio Murino, Alessio Del Bue

To this end, we propose a novel approach that applies a guided attention mechanism between the objects, and the spatiotemporal features extracted from video clips, enhancing the motion and contextual information, and further decoding the object-centric and motion-centric information to address the problem of STA in egocentric videos.

Action Anticipation Object +1

Target-driven One-Shot Unsupervised Domain Adaptation

no code implementations8 May 2023 Julio Ivan Davila Carrazco, Suvarna Kishorkumar Kadam, Pietro Morerio, Alessio Del Bue, Vittorio Murino

Unlike existing methods, our augmentation module allows for strong transformations of the source samples, and the style of the single target sample available is exploited to guide the augmentation by ensuring perceptual similarity.

One-shot Unsupervised Domain Adaptation Unsupervised Domain Adaptation

Continual Source-Free Unsupervised Domain Adaptation

no code implementations14 Apr 2023 Waqar Ahmed, Pietro Morerio, Vittorio Murino

Existing Source-free Unsupervised Domain Adaptation (SUDA) approaches inherently exhibit catastrophic forgetting.

Continual Learning Unsupervised Domain Adaptation

Guiding Pseudo-labels with Uncertainty Estimation for Source-free Unsupervised Domain Adaptation

2 code implementations CVPR 2023 Mattia Litrico, Alessio Del Bue, Pietro Morerio

We propose a novel approach for the SF-UDA setting based on a loss reweighting strategy that brings robustness against the noise that inevitably affects the pseudo-labels.

Unsupervised Domain Adaptation

Self-improving object detection via disagreement reconciliation

no code implementations21 Feb 2023 Gianluca Scarpellini, Stefano Rosa, Pietro Morerio, Lorenzo Natale, Alessio Del Bue

Object detectors often experience a drop in performance when new environmental conditions are insufficiently represented in the training data.

Object object-detection +1

Anticipating Next Active Objects for Egocentric Videos

no code implementations13 Feb 2023 Sanket Thakur, Cigdem Beyan, Pietro Morerio, Vittorio Murino, Alessio Del Bue

This paper addresses the problem of anticipating the next-active-object location in the future, for a given egocentric video clip where the contact might happen, before any action takes place.

Object

Look around and learn: self-improving object detection by exploration

no code implementations7 Feb 2023 Gianluca Scarpellini, Stefano Rosa, Pietro Morerio, Lorenzo Natale, Alessio Del Bue

Object detectors often experience a drop in performance when new environmental conditions are insufficiently represented in the training data.

Object object-detection +1

3DSGrasp: 3D Shape-Completion for Robotic Grasp

no code implementations2 Jan 2023 Seyed S. Mohammadi, Nuno F. Duarte, Dimitris Dimou, Yiming Wang, Matteo Taiana, Pietro Morerio, Atabak Dehban, Plinio Moreno, Alexandre Bernardino, Alessio Del Bue, Jose Santos-Victor

However, in practice, PCDs are often incomplete when objects are viewed from few and sparse viewpoints before the grasping action, leading to the generation of wrong or inaccurate grasp poses.

Robotic Grasping

Compact CNN Structure Learning by Knowledge Distillation

no code implementations19 Apr 2021 Waqar Ahmed, Andrea Zunino, Pietro Morerio, Vittorio Murino

The concept of compressing deep Convolutional Neural Networks (CNNs) is essential to use limited computation, power, and memory resources on embedded devices.

Knowledge Distillation Model Compression

Adaptive Pseudo-Label Refinement by Negative Ensemble Learning for Source-Free Unsupervised Domain Adaptation

no code implementations29 Mar 2021 Waqar Ahmed, Pietro Morerio, Vittorio Murino

On the contrary, a pre-trained source model is always considered to be available, even though performing poorly on target due to the well-known domain shift problem.

Ensemble Learning Pseudo Label +1

Single Image Human Proxemics Estimation for Visual Social Distancing

1 code implementation3 Nov 2020 Maya Aghaei, Matteo Bustreo, Yiming Wang, Gianluca Bailo, Pietro Morerio, Alessio Del Bue

In this work, we address the problem of estimating the so-called "Social Distancing" given a single uncalibrated image in unconstrained scenarios.

Cross-modal Speaker Verification and Recognition: A Multilingual Perspective

no code implementations28 Apr 2020 Muhammad Saad Saeed, Shah Nawaz, Pietro Morerio, Arif Mahmood, Ignazio Gallo, Muhammad Haroon Yousaf, Alessio Del Bue

Recent years have seen a surge in finding association between faces and voices within a cross-modal biometric application along with speaker recognition.

Speaker Recognition Speaker Verification

Intra-Camera Supervised Person Re-Identification

no code implementations12 Feb 2020 Xiangping Zhu, Xiatian Zhu, Minxian Li, Pietro Morerio, Vittorio Murino, Shaogang Gong

Existing person re-identification (re-id) methods mostly exploit a large set of cross-camera identity labelled training data.

Person Re-Identification

Generative Pseudo-label Refinement for Unsupervised Domain Adaptation

2 code implementations9 Jan 2020 Pietro Morerio, Riccardo Volpi, Ruggero Ragonesi, Vittorio Murino

We exploit this finding in an iterative procedure where a generative model and a classifier are jointly trained: in turn, the generator allows to sample cleaner data from the target distribution, and the classifier allows to associate better labels to target samples, progressively refining target pseudo-labels.

Pseudo Label Unsupervised Domain Adaptation

DMCL: Distillation Multiple Choice Learning for Multimodal Action Recognition

1 code implementation23 Dec 2019 Nuno C. Garcia, Sarah Adel Bargal, Vitaly Ablavsky, Pietro Morerio, Vittorio Murino, Stan Sclaroff

In this work, we address the problem of learning an ensemble of specialist networks using multimodal data, while considering the realistic and challenging scenario of possible missing modalities at test time.

Action Recognition Multiple-choice +1

Unsupervised Domain-Adaptive Person Re-identification Based on Attributes

no code implementations27 Aug 2019 Xiangping Zhu, Pietro Morerio, Vittorio Murino

Training person re-identification (ReID) algorithms under the supervision of such attributes have proven to be effective in extracting local features which are important for ReID.

Attribute Domain Adaptation +3

Audio-Visual Model Distillation Using Acoustic Images

1 code implementation16 Apr 2019 Andrés F. Pérez, Valentina Sanguineti, Pietro Morerio, Vittorio Murino

In this paper, we investigate how to learn rich and robust feature representations for audio classification from visual data and acoustic images, a novel audio data modality.

Action Recognition Audio Classification +3

Cross-modal Learning by Hallucinating Missing Modalities in RGB-D Vision

1 code implementation Multimodal Scene Understanding Algorithms, Applications and Deep Learning 2019 Nuno C. Garcia, Pietro Morerio, Vittorio Murino

We report state-of-the-art or comparable results on video action recognition on the largest multimodal dataset available for this task, the NTU RGB+D, as well as on the UWA3DII and Northwestern-UCLA.

Action Recognition Hallucination +3

Learning with privileged information via adversarial discriminative modality distillation

1 code implementation19 Oct 2018 Nuno C. Garcia, Pietro Morerio, Vittorio Murino

This raises the challenge of how to extract information from multimodal data in the training stage, in a form that can be exploited at test time, considering limitations such as noisy or missing modalities.

Action Recognition Hallucination

Modality Distillation with Multiple Stream Networks for Action Recognition

1 code implementation ECCV 2018 Nuno Garcia, Pietro Morerio, Vittorio Murino

Particularly, we consider the case of learning representations from depth and RGB videos, while relying on RGB data only at test time.

Action Classification Action Detection +3

Excitation Dropout: Encouraging Plasticity in Deep Neural Networks

1 code implementation23 May 2018 Andrea Zunino, Sarah Adel Bargal, Pietro Morerio, Jianming Zhang, Stan Sclaroff, Vittorio Murino

In this work, we utilize the evidence at each neuron to determine the probability of dropout, rather than dropping out neurons uniformly at random as in standard dropout.

Decision Making Video Recognition

Scalable and Compact 3D Action Recognition with Approximated RBF Kernel Machines

no code implementations28 Nov 2017 Jacopo Cavazza, Pietro Morerio, Vittorio Murino

Despite the recent deep learning (DL) revolution, kernel machines still remain powerful methods for action recognition.

3D Action Recognition

Minimal-Entropy Correlation Alignment for Unsupervised Deep Domain Adaptation

1 code implementation ICLR 2018 Pietro Morerio, Jacopo Cavazza, Vittorio Murino

In this work, we face the problem of unsupervised domain adaptation with a novel deep learning approach which leverages on our finding that entropy minimization is induced by the optimal alignment of second order statistics between source and target domains.

General Classification Unsupervised Domain Adaptation

Adversarial Feature Augmentation for Unsupervised Domain Adaptation

2 code implementations CVPR 2018 Riccardo Volpi, Pietro Morerio, Silvio Savarese, Vittorio Murino

Recent works showed that Generative Adversarial Networks (GANs) can be successfully applied in unsupervised domain adaptation, where, given a labeled source dataset and an unlabeled target dataset, the goal is to train powerful classifiers for the target samples.

Data Augmentation Unsupervised Domain Adaptation

Dropout as a Low-Rank Regularizer for Matrix Factorization

no code implementations13 Oct 2017 Jacopo Cavazza, Pietro Morerio, Benjamin Haeffele, Connor Lane, Vittorio Murino, Rene Vidal

Regularization for matrix factorization (MF) and approximation problems has been carried out in many different ways.

A Compact Kernel Approximation for 3D Action Recognition

no code implementations6 Sep 2017 Jacopo Cavazza, Pietro Morerio, Vittorio Murino

In this work we reduce such complexity to be linear by proposing a novel and explicit feature map to approximate the kernel function.

3D Action Recognition

When Kernel Methods meet Feature Learning: Log-Covariance Network for Action Recognition from Skeletal Data

no code implementations3 Aug 2017 Jacopo Cavazza, Pietro Morerio, Vittorio Murino

Human action recognition from skeletal data is a hot research topic and important in many open domain applications of computer vision, thanks to recently introduced 3D sensors.

Action Recognition Temporal Action Localization

Correlation Alignment by Riemannian Metric for Domain Adaptation

1 code implementation23 May 2017 Pietro Morerio, Vittorio Murino

Domain adaptation techniques address the problem of reducing the sensitivity of machine learning methods to the so-called domain shift, namely the difference between source (training) and target (test) data distributions.

Unsupervised Domain Adaptation

Curriculum Dropout

2 code implementations ICCV 2017 Pietro Morerio, Jacopo Cavazza, Riccardo Volpi, Rene Vidal, Vittorio Murino

This induces an adaptive regularization scheme that smoothly increases the difficulty of the optimization problem.

Image Classification Scheduling

Left/Right Hand Segmentation in Egocentric Videos

no code implementations21 Jul 2016 Alejandro Betancourt, Pietro Morerio, Emilia Barakova, Lucio Marcenaro, Matthias Rauterberg, Carlo Regazzoni

Due to their favorable location, wearable cameras frequently capture the hands of the user, and may thus represent a promising user-machine interaction tool for different applications.

Hand Segmentation Segmentation +1

The Evolution of First Person Vision Methods: A Survey

no code implementations4 Sep 2014 Alejandro Betancourt, Pietro Morerio, Carlo S. Regazzoni, Matthias Rauterberg

The emergence of new wearable technologies such as action cameras and smart-glasses has increased the interest of computer vision scientists in the First Person perspective.

Activity Recognition object-detection +1

Cannot find the paper you are looking for? You can Submit a new open access paper.