知搜-全球专利检索

  1. Login
  2. Sign Up

Search Results

1259 records (took 0.053 seconds)

    INFORMATION PROCESSING APPARATUS, INFORMATION PROCESSING METHOD, AND STORAGE MEDIUM
    71.
    发明公开
    INFORMATION PROCESSING APPARATUS, INFORMATION PROCESSING METHOD, AND STORAGE MEDIUM 审中-公开

    公开(公告)号：US20240163558A1

    公开(公告)日：2024-05-16

    申请号：US18495849

    申请日：2023-10-27

    Applicant: CANON KABUSHIKI KAISHA

    Inventor： TAKUYA TOYODA

    IPC: H04N23/69 ,  G06T7/20 ,  G06V10/25 ,  H04N23/61

    CPC classification number: H04N23/69 ,  G06T7/20 ,  G06V10/25 ,  H04N23/61 ,  G06V2201/07

    Abstract: An information processing apparatus includes a detection unit configured to detect a predetermined target object from an image that has been captured by an image capturing unit, a setting unit configured to set a predetermined region with respect to the image, a determination unit configured to perform a determination as to whether or not the target object has exited from the inside of the predetermined region, a control unit configured to control an image capturing range of the image capturing unit according to a result that has been determined by the determination unit, and a zoom magnification setting unit configured to set the zoom magnification of the image capturing unit to the same zoom magnification as that before the target object enters the predetermined region, in a case in which the determination unit determines that the target object has exited from the inside of the predetermined region.

    AUGMENTED REALITY METHOD FOR MONITORING AN EVENT IN A SPACE COMPRISING AN EVENT FIELD IN REAL TIME
    78.
    发明公开
    AUGMENTED REALITY METHOD FOR MONITORING AN EVENT IN A SPACE COMPRISING AN EVENT FIELD IN REAL TIME 审中-公开

    公开(公告)号：US20240144613A1

    公开(公告)日：2024-05-02

    申请号：US18406105

    申请日：2024-01-06

    Applicant: IMMERSIV

    Inventor： STÉPHANE GUERIN ,  EMMANUELLE ROGER

    IPC: G06T19/00 ,  G06V10/74 ,  G06V20/52

    CPC classification number: G06T19/006 ,  G06V10/761 ,  G06V20/52 ,  G06V2201/07

    Abstract: An augmented reality method for monitoring an event in space includes acquisition of a plurality of images of space having at least two landmarks by a camera of a portable device. Space being associated with a three-dimensional reference frame and the portable device being associated with a two-dimensional reference frame. A three-dimensional position and orientation of the space in relation to the camera is determined. The instantaneous position, within the reference frame of the space, of a mobile element moving in the space is received. The position of the mobile element in the two-dimensional reference frame is calculated from transformation parameters calculated from the three-dimensional position and orientation of the space in relation to the camera. An overlay at a predetermined distance in relation to the position of the mobile element in the two-dimensional reference frame is displayed on the screen. Also, a portable electronic device implements the method.

    METHODS AND SYSTEMS FOR DETERMINING 3D COORDINATES DESCRIBING A SCENE
    79.
    发明公开
    METHODS AND SYSTEMS FOR DETERMINING 3D COORDINATES DESCRIBING A SCENE 审中-公开

    公开(公告)号：US20240144509A1

    公开(公告)日：2024-05-02

    申请号：US18067657

    申请日：2022-12-16

    Applicant: Maxar International Sweden AB

    Inventor： Ola NYGREN ,  Folke ISAKSSON

    IPC: G06T7/60 ,  G06T7/80 ,  G06V10/74 ,  G06V20/13

    CPC classification number: G06T7/60 ,  G06T7/80 ,  G06V10/761 ,  G06V20/13 ,  G06V2201/07

    Abstract: The disclosure relates to a computer implemented method for determining 3D coordinates describing a scene, obtaining first, second and third images covering the scene captured from different viewpoints, said images being associated to a timing of capture, wherein the scene is formed in an overlapping geographical area of the first, second and third images. The method further comprises obtaining intrinsic and/or extrinsic parameter values for at least one imaging device used for capture of the images, and determining the 3D coordinates describing the scene, said determination comprising performing bundle-adjustments based on pair-wise measurements on the images to minimize a re-projection error of reconstructed 3D points in the overlapping geographical area, wherein time is used as a parameter in the determination of the 3D coordinates describing the scene, and wherein the re-projection error is minimized allowing at least a height to be locally time dependent in reconstruction of 3D points at least for a part of the overlapping geographical area.

    METHOD FOR LEVERAGING SHAPE INFORMATION IN FEW-SHOT LEARNING IN ARTIFICIAL NEURAL NETWORKS
    80.
    发明公开
    METHOD FOR LEVERAGING SHAPE INFORMATION IN FEW-SHOT LEARNING IN ARTIFICIAL NEURAL NETWORKS 审中-公开

    公开(公告)号：US20240135722A1

    公开(公告)日：2024-04-25

    申请号：US18165857

    申请日：2023-02-07

    Applicant: NavInfo Europe B.V.

    Inventor： Deepan Chakravarthi Padmanabhan ,  Shruthi Gowda ,  Elahe Arani ,  Bahram Zonooz

    IPC: G06V20/58 ,  G06V10/82 ,  G06V20/40

    CPC classification number: G06V20/58 ,  G06V10/82 ,  G06V20/41 ,  G06V2201/07

    Abstract: A computer-implemented method that provides a novel shape aware FSL framework, referred to as LSFSL. In addition to the inductive biases associated with deep learning models, the method of the current invention introduces meaningful shape bias. The method of the current invention comprises the step of capturing the human behavior of recognizing objects by utilizing shape information. The shape information is distilled to address the texture bias of CNN-based models. During training, the model has two branches: RIN-branch, network with colored images as input, preferably RGB images, and SIN-branch, network with shape semantic-based input. Each branch incorporates a CNN backbone followed by a fully connected layer performing classification. RIN-branch and SIN-branch receive the RGB input image and shape information enhanced RGB input image, respectively. The training objective is to improve the classification performance of the RIN-branch and SIN-branch as well as to distill shape semantics from SIN-branch to RIN-branch. The features of the RIN-branch and SIN-branch are aligned to distill shape representation into RIN-branch. This feature alignment implicitly achieves a bias-alignment between the RIN and SIN. The learned representations are generic and remain invariant to common attributes.

Patent Agency Ranking