E
Elisa Ricci
Researcher at fondazione bruno kessler
Publications - 44
Citations - 2159
Elisa Ricci is an academic researcher from fondazione bruno kessler. The author has contributed to research in topics: Deep learning & Convolutional neural network. The author has an hindex of 20, co-authored 44 publications receiving 1342 citations. Previous affiliations of Elisa Ricci include Sapienza University of Rome & University of Trento.
Papers
More filters
Proceedings ArticleDOI
Animating Arbitrary Objects via Deep Motion Transfer
TL;DR: In this article, a deep learning framework for image animation and video generation is proposed, which consists of a keypoint detector unsupervisely trained to extract object keypoints, a dense motion prediction network for generating dense heatmaps from sparse keypoints and a motion transfer network for synthesizing the output frames.
Proceedings ArticleDOI
Unsupervised Domain Adaptation Using Feature-Whitening and Consensus Loss
TL;DR: In this paper, the authors propose domain alignment layers which implement feature whitening for the purpose of matching source and target feature distributions, and leverage the unlabeled target data by proposing the Min-Entropy Consensus loss, which regularizes training while avoiding the adoption of many user-defined hyper-parameters.
Proceedings ArticleDOI
Unsupervised Adversarial Depth Estimation Using Cycled Generative Networks
TL;DR: Pilzer et al. as mentioned in this paper propose a deep generative network that learns to predict the correspondence field (i.e., the disparity map) between two image views in a calibrated stereo camera setting.
Proceedings ArticleDOI
Boosting Domain Adaptation by Discovering Latent Domains
TL;DR: In this article, the authors propose a novel Convolutional neural network (CNN) architecture which automatically discovers latent domains in visual datasets and exploits this information to learn robust target classifiers.
Proceedings ArticleDOI
Best Sources Forward: Domain Generalization through Source-Specific Nets
TL;DR: In this paper, a deep network with multiple domain-specific classifiers, each associated to a source domain, is designed to learn domain invariant representations from source data and exploit the probabilities that a target sample belongs to each source domain and exploit them to optimally fuse the classifiers predictions.