公开(公告)号：US20210326376A1

公开(公告)日：2021-10-21

申请号：US17313545

申请日：2021-05-06

Applicant: Facebook, Inc.

Inventor： Jan Erik Solem ,  Yubin Kuang ,  Johan Gyllenspetz ,  Peter Neubauer ,  Pau Gargallo Piracés

IPC: G06F16/54 ,  G06F16/29 ,  G06F16/9537 ,  G06F16/58 ,  G01C21/36 ,  G06F16/583 ,  G06F16/51 ,  G06F16/2457

Abstract: Information relating to a subset of images within a set of images is received. The information includes metadata of each image. The subset of images are selected from the set of images based upon each image in the subset of images being acquired within a bounding area surrounding a geographical position of a current image. Weight information is received. The weight information controls weights to be given to a plurality of different measures of similarities. A plurality of combined measures of similarity for pairs of images in the subset of images are calculated.

公开(公告)号：US20220198710A1

公开(公告)日：2022-06-23

申请号：US17439355

申请日：2020-03-13

Applicant: Facebook, Inc.

Inventor： Yubin Kuang ,  Pau Gargallo ,  Manuel López Antequera ,  Roger Marí Molas ,  Jan Erik Solem

IPC: G06T7/80 ,  G06T5/00

Abstract: A method for analysis of an image comprises: receiving (402) the image to be analyzed; processing (404) the image with a machine-learned model, wherein the machine-learned model is configured to predict at least an intrinsic parameter of the image using at least a first variable of the machine-learned model, wherein the first variable defines a relation between a radial distortion of the image and a focal length of the image; and outputting (406) the intrinsic parameter of the image. Also, methods for forming a 3D reconstruction of a scenery, for training a machine-learned model for analysis of an image and for generating a dataset of images for training a machine-learned model are disclosed.

公开(公告)号：US11030235B2

公开(公告)日：2021-06-08

申请号：US16066719

申请日：2016-12-27

Applicant: Facebook, Inc.

Inventor： Jan Erik Solem ,  Yubin Kuang ,  Johan Gyllenspetz ,  Peter Neubauer ,  Pau Gargallo Piracés

IPC: G06F16/54 ,  G06F16/29 ,  G06F16/9537 ,  G06F16/58 ,  G01C21/36 ,  G06F16/583 ,  G06F16/51 ,  G06F16/2457

Abstract: The present invention relates to a method for navigating through a set of images depicting a scenery, said method comprising: receiving information relating to a subset of images within said set of images, said information comprising metadata of each image, said subset of images comprising images being acquired from a geographical position, which is within a bounding area surrounding the geographical position of a current image within the set of images; receiving weight information, said weight information controlling weights to be given to a plurality of different measures of similarities based on different types of metadata; and calculating a plurality of combined measures of similarity for pairs of images in the subset of images, said calculating being performed using said metadata and said weight information.