Picture for Mariella Dimiccoli

Mariella Dimiccoli

3M-TRANSFORMER: A Multi-Stage Multi-Stream Multimodal Transformer for Embodied Turn-Taking Prediction

Add code
Nov 01, 2023
Viaarxiv icon

Debiased-CAM to mitigate systematic error with faithful visual explanations of machine learning

Add code
Jan 30, 2022
Viaarxiv icon

Enhancing Egocentric 3D Pose Estimation with Third Person Views

Add code
Jan 07, 2022
Figure 1 for Enhancing Egocentric 3D Pose Estimation with Third Person Views
Figure 2 for Enhancing Egocentric 3D Pose Estimation with Third Person Views
Figure 3 for Enhancing Egocentric 3D Pose Estimation with Third Person Views
Figure 4 for Enhancing Egocentric 3D Pose Estimation with Third Person Views
Viaarxiv icon

Learning grounded word meaning representations on similarity graphs

Add code
Sep 07, 2021
Figure 1 for Learning grounded word meaning representations on similarity graphs
Figure 2 for Learning grounded word meaning representations on similarity graphs
Figure 3 for Learning grounded word meaning representations on similarity graphs
Figure 4 for Learning grounded word meaning representations on similarity graphs
Viaarxiv icon

Graph Constrained Data Representation Learning for Human Motion Segmentation

Add code
Jul 28, 2021
Figure 1 for Graph Constrained Data Representation Learning for Human Motion Segmentation
Figure 2 for Graph Constrained Data Representation Learning for Human Motion Segmentation
Figure 3 for Graph Constrained Data Representation Learning for Human Motion Segmentation
Figure 4 for Graph Constrained Data Representation Learning for Human Motion Segmentation
Viaarxiv icon

Interaction-GCN: a Graph Convolutional Network based framework for social interaction recognition in egocentric videos

Add code
Apr 28, 2021
Figure 1 for Interaction-GCN: a Graph Convolutional Network based framework for social interaction recognition in egocentric videos
Figure 2 for Interaction-GCN: a Graph Convolutional Network based framework for social interaction recognition in egocentric videos
Figure 3 for Interaction-GCN: a Graph Convolutional Network based framework for social interaction recognition in egocentric videos
Figure 4 for Interaction-GCN: a Graph Convolutional Network based framework for social interaction recognition in egocentric videos
Viaarxiv icon

Modeling long-term interactions to enhance action recognition

Add code
Apr 23, 2021
Figure 1 for Modeling long-term interactions to enhance action recognition
Figure 2 for Modeling long-term interactions to enhance action recognition
Figure 3 for Modeling long-term interactions to enhance action recognition
Figure 4 for Modeling long-term interactions to enhance action recognition
Viaarxiv icon

Debiased-CAM for bias-agnostic faithful visual explanations of deep convolutional networks

Add code
Dec 10, 2020
Figure 1 for Debiased-CAM for bias-agnostic faithful visual explanations of deep convolutional networks
Figure 2 for Debiased-CAM for bias-agnostic faithful visual explanations of deep convolutional networks
Figure 3 for Debiased-CAM for bias-agnostic faithful visual explanations of deep convolutional networks
Figure 4 for Debiased-CAM for bias-agnostic faithful visual explanations of deep convolutional networks
Viaarxiv icon

Seeing and Hearing Egocentric Actions: How Much Can We Learn?

Add code
Oct 15, 2019
Figure 1 for Seeing and Hearing Egocentric Actions: How Much Can We Learn?
Figure 2 for Seeing and Hearing Egocentric Actions: How Much Can We Learn?
Figure 3 for Seeing and Hearing Egocentric Actions: How Much Can We Learn?
Figure 4 for Seeing and Hearing Egocentric Actions: How Much Can We Learn?
Viaarxiv icon

Learning event representations in image sequences by dynamic graph embedding

Add code
Oct 08, 2019
Figure 1 for Learning event representations in image sequences by dynamic graph embedding
Figure 2 for Learning event representations in image sequences by dynamic graph embedding
Figure 3 for Learning event representations in image sequences by dynamic graph embedding
Figure 4 for Learning event representations in image sequences by dynamic graph embedding
Viaarxiv icon