Self adaptive matrix completion for heart rate estimation from face videos under realistic conditions
Department/Centre: Department of Information Engineering and Computer Science
Inventors: Jeffrey Cohn (University of Pittsburgh), Niculae Sebe (UniTrento), Xavier Alameda-Pineda (UniTrento), Sergey Tulyakov (UniTrento), Elisa Ricci (Fondazione Bruno Kessler), Lijun Yin (University of Binghamton)
Field of use: heart beat recognition
Keywords: video analysis, surveillance
Bibliographical data: application number: 15/631,346 (US filing); priority date: 23/6/2017; patent granted in USA (US 10,335,045) on 2/7/2019
Assignees: Università di Trento, University of Binghamton, University of Pittsburgh, Fondazione Bruno Kessler
Availability: available for license or assignment
Description
This invention, inspired by recent advances on matrix completion theory, allows to predict the HR while simultaneously discovering the best regions of the face to be used for estimation.
Recent studies in computer vision have shown that, while practically invisible to a human observer, skin color changes due to blood flow can be captured on face videos and, surprisingly, be used to estimate the heart rate (HR). While considerable progress has been made in the last few years, still many issues remain open. In particular, state-of-the-art approaches are not robust enough to operate in natural conditions (e.g. in case of spontaneous movements, facial expressions or illumination changes).
Opposite to previous approaches that estimate the HR by processing all the skin pixels inside a fixed region of interest, the new approach dynamically selects face regions useful for robust HR estimation.
Method for the computer-implemented generation of a synthetic data set for training a convolutional neural network for an Interferometric SAR
Department/Centre: Department of Information Engineering and Computer Science
Inventors: Lorenzo Bruzzone (UniTrento); Francescopaolo Sica, Giorgia Gobbi, Paola Rizzoli (Deutsches Zentrum für Luft- und Raumfahrt)
Field of use: high-resolution InSAR processing and Digital Elevation Model (DEM) generation, differential Interferometry (D-InSAR) and Persistent Scatterers (PS), subsidence and tectonic monitoring, earthquakes monitoring, convolutional Neural Networks.
Keywords: Synthetic Aperture Radar (SAR), Interferometry, Deep Learning, Machine Learning, Artificial Intelligence, Remote Sensing.
Bibliographical data: application numer: 20169572.3 (EU filing); priority date: 15/04/2020
Assignees: Università di Trento, Deutsches Zentrum für Luft- und Raumfahrt
Availability: available for license or assignment
Descrizione
This invention describes a method for the computer-implemented generation of a synthetic data set for training a convolutional neural network for an Interferometric Synthetic Aperture Radar.
Non-volatile intermittent processing on FPGA
Title: Non-volatile intermittent processing on FPGA
Department/Centre: Department of Industrial Engineering / Department of Information Engineering and Computer Science
Inventors: Davide Brunelli (UniTrento), Kasim Sinan Yildirim (UniTrento)
Field of use: verify the behavior of typical of transiently-powered computing systems equipped with non-volatile memories; fast-prototype and debug non-volatile processor architectures without the production of pre-industrial hardware prototypes.
Keywords: Intermittent Computing, Transiently-powered Computers, Non-volatile Processors, Non-volatile Logic, FPGAs
Bibliographical data: application number: 102020000022114 (Italian filing); priority date: 18/09/2020
Assignees: Università di Trento
Availability: available for license or assignment
Description
This invention refers to a new FPGA architecture that enables the emulation and behavioral verification of the computing systems exploiting fast non-volatile memories.