公开(公告)号：US20230114215A1

公开(公告)日：2023-04-13

申请号：US18045824

申请日：2022-10-11

Applicant: AUTOBRAINS TECHNOLOGIES LTD

Inventor： Tom TABAK

IPC: G06V20/56 ,  G06V10/762 ,  G06V10/774

Abstract: A method for lane boundary detection, the method may include obtaining an image of an environment of a vehicle, the environment comprises at least one lane boundary portions; wherein the image comprises a first plurality of two dimensional (2D) image segments; converting, by a first machine learning process, each of the 2D segments to a segment vector to provide a first plurality of segment vectors; wherein each segment vector represents a 2D segment; finding an associated cluster for each segment vector to provide a second plurality of associated clusters; searching for at least one LBR cluster of the second plurality of associated clusters; and determining, for each LBR segment vector and by a second machine learning process, a location of a lane boundary portion within a 2D image segment that is represented by the LBR segment vector; wherein a LBR segment vector has an associated cluster that is a LBR cluster.

公开(公告)号：US20240075925A1

公开(公告)日：2024-03-07

申请号：US18461480

申请日：2023-09-05

Applicant: AUTOBRAINS TECHNOLOGIES LTD

Inventor： JONATHAN COHEN ,  Tom TABAK

IPC: B60W30/12 ,  B60W10/20 ,  B60W40/072 ,  B60W60/00 ,  G06V10/82 ,  G06V20/56

CPC classification number: B60W30/12 ,  B60W10/20 ,  B60W40/072 ,  B60W60/001 ,  G06V10/82 ,  G06V20/588 ,  B60W2420/42 ,  B60W2552/30 ,  B60W2552/53 ,  G06V2201/10

Abstract: A method for steering assistance, the method may include obtaining by vehicle camera, an image of an environment of the vehicle; feeding the image to a machine learning process; outputting, by the machine learning process, road line metadata that comprises a line offset value, a line angle value, a line curvature value and a line curvature rate; and preforming a steering related response to the road line metadata.

公开(公告)号：US20230119374A1

公开(公告)日：2023-04-20

申请号：US18045820

申请日：2022-10-11

Applicant: AUTOBRAINS TECHNOLOGIES LTD

Inventor： Tom TABAK

IPC: G06T7/10 ,  G06N3/08

Abstract: A method for neural network based image processing, the method may include obtaining multiple two dimensional (2D) segments of a sensed information unit, each 2D segment has a segment location within the sensed information unit; converting each 2D segments to a segment vector; generating multiple segments intermediate results by repeating, for each segment vector: (a) concatenating the segment vector with associated segment location information to provide a first segment concatenated vector; and (b) multiplying the first segment concatenated vector by a learnable embedding matrix to provide a segment intermediate result; concatenating the multiple segments intermediate results with associated segment identifiers to provide a sensed information unit result; and feeding the sensed information unit result to a second layer of a neural network.

公开(公告)号：US20230045885A1

公开(公告)日：2023-02-16

申请号：US17805847

申请日：2022-06-07

Applicant: AUTOBRAINS TECHNOLOGIES LTD

Inventor： Tom TABAK

IPC: G06V20/56 ,  G06V10/82

Abstract: A method for context based lane prediction, the method may include obtaining sensed information regarding an environment of the vehicle; providing the sensed information to a second trained machine learning process; and locating one or more lane boundaries by the second trained machine learning process. The second trained machine learning process is generated by: performing a self-supervised training process, using a first dataset, of a first machine learning process to provide a first trained machine learning process; wherein the first trained machine learning process comprises a first encoder portion and a first decoder portion; replacing the first decoder portion by a second decoder portion to provide a second machine learning process; and performing an additional training process, using a second dataset that is associated with lane boundary metadata, of the second machine learning process to provide a second trained machine learning process.