University of Michigan

Carnegie Mellon University

University of Rochester

Walter S Lasecki

Christopher Harrison

Gierad Laput

Jeffrey P Bigham

Bo Robert Xiao

Stephan J Lemmer

Jason J Corso

Jean Young S Kwon

Jason Wiese

A method is disclosed for reconstructing three-dimensional video from two-dimensional video data using particle filtering and thereby generating training data for autonomous vehicles. In one version, the method comprises: receiving a set of annotations associated with a video frame comprising a view of at least a portion of a vehicle, each annotation comprising at least one two-dimensional line; removing at least one outlier from the set of annotations; determining an estimated vehicle model based on the set of annotations; and providing the estimated vehicle model to a driving simulator.

Reconstructing 3D video using particle filtering to aggregate crowd responses

The disclosure describes a sensor system that provides end users with intelligent sensing capabilities, and embodies both crowd sourcing and machine learning together. Further, a sporadic crowd assessment is used to ensure continued sensor accuracy when the system is relying on machine learning analysis. This sensor approach requires minimal and non-permanent sensor installation by utilizing any device with a camera as a sensor host, and provides human-centered and actionable sensor output.

System and method for adaptive, rapidly deployable, human-intelligent sensor feeds

Growing community of inventors

Ann Arbor, MI, United States of America

Walter S Lasecki