Verteilbares Darstellungslernen zum Verknüpfen von Beobachtungen aus mehreren Fahrzeugen

Details

Veröffentlichungs-Nr.

EP3511863

Aktenzeichen

EP19151349

Anmeldetag

11. Januar 2019

Veröffentlichung

2. August 2023

Erteilung

2. August 2023

Rechtsraum

EP

IPC

G06V20/56G06V10/764G06V10/771G06F18/2113G06F18/2413

Abstract

In an example embodiment, a computer-implemented method generates, using first machine learning logic, a first compact representation of a first object in a first image captured from a first perspective; generates, using second machine learning logic, a second compact representation of a second object in a second image captured from a second perspective; computes a similarity score reflecting a level of similarity between the first compact representation of the first object and the second compact representation of the second object; and identifies a subset of features of the first object and the second object as being more determinative than other features of the first object and the second object based on a comparison of the similarity score to a predetermined target output.

Anmelder

Firma

TOYOTA JIDOSHA KABUSHIKI KAISHA

Land

🇯🇵 Japan

Fachgebiete