S
Stefano Soatto
Researcher at University of California, Los Angeles
Publications - 499
Citations - 27815
Stefano Soatto is an academic researcher from University of California, Los Angeles. The author has contributed to research in topics: Motion estimation & Image segmentation. The author has an hindex of 78, co-authored 499 publications receiving 23597 citations. Previous affiliations of Stefano Soatto include University of California & University of California, Davis.
Papers
More filters
Posted Content
Learning Topology from Synthetic Data for Unsupervised Depth Completion
TL;DR: This work presents a method for inferring dense depth maps from images and sparse depth measurements by leveraging synthetic data to learn the association of sparse point clouds with dense natural shapes, and using the image as evidence to validate the predicted depth map.
Proceedings ArticleDOI
Learning and matching multiscale template descriptors for real-time detection, localization and tracking
Taehee Lee,Stefano Soatto +1 more
TL;DR: A system to learn an object template from a video stream, and localize and track the corresponding object in live video, thus enabling detection and tracking in spite of partial occlusion.
Proceedings ArticleDOI
Multi-view feature engineering and learning
Jingming Dong,Nikolaos Karianakis,Damek Davis,Joshua Hernandez,Jonathan Balzer,Stefano Soatto +5 more
TL;DR: A sampling-based and a point-estimate based approximation of a local representation of imaging data are proposed, compared empirically on image-to-(multiple)image matching, and a multi-view wide-baseline matching benchmark is introduced.
Proceedings ArticleDOI
Learning Semantic-Aware Dynamics for Video Prediction
TL;DR: In this paper, the authors propose an architecture and training scheme to predict video frames by explicitly modeling dis-occlusions and capturing the evolution of semantically consistent regions in the video, where scene layout and motion are decomposed into layers, which are predicted and fused with their context to generate future layouts and motions.
Posted Content
On the energy landscape of deep networks
Pratik Chaudhari,Stefano Soatto +1 more
TL;DR: It is shown that a regularization term akin to a magnetic field can be modulated with a single scalar parameter to transition the loss function from a complex, non-convex landscape with exponentially many local minima, to a phase with a polynomial number of minima and all the way down to a trivial landscape with a unique minimum.