公开(公告)号：US12229926B2

公开(公告)日：2025-02-18

申请号：US17863501

申请日：2022-07-13

Applicant: GM GLOBAL TECHNOLOGY OPERATIONS LLC

Inventor： Tzvi Philipp ,  Dan Levi ,  Shai Silberstein

IPC: G06T5/73

Abstract: A method for deblurring a blurred image includes dividing the blurred image into overlapping regions each having a size and an offset from neighboring overlapping regions along a first direction as determined by a period of a ringing artifact in the blurred image, or by obtained blur characteristics relating to the blurred image and/or attributable to the optical system, or by a detected cause capable of producing the blur characteristics, stacking the overlapping regions to produce a stacked output, wherein the overlapping regions are sequentially organized along the first direction, convolving the stacked output through a first convolutional neural network (CNN) to produce a first CNN output having reduced blur as compared to the stacked output, and assembling the first CNN output into a re-assembled image, and processing the re-assembled image through a second CNN to produce a deblurred image having reduced residual artifacts as compared to the re-assembled image.

公开(公告)号：US10046715B2

公开(公告)日：2018-08-14

申请号：US14918267

申请日：2015-10-20

Applicant: GM GLOBAL TECHNOLOGY OPERATIONS LLC

Inventor： Dan Levi ,  Shai Silberstein

IPC: G06K9/00 ,  B60R11/04 ,  H04N7/18 ,  B60R1/00

Abstract: Systems and methods for detecting objects from a moving platform are provided. The method includes receiving, from a camera coupled to the moving platform, first and second image frames, wherein the second image frame is subsequent in time to the first image frame. Within the first image frame, a first target is defined characterizing the location of an object detected in the first image frame. A first score associated with the first target and a second score associated with the second target are determined, and a first predicted target is determined based on a tracking process applied to the first target. A third score associated with the first predicted target is determined. The method determines a merged target corresponding to the second target when the second score is greater than the third score and a proximity measure between the second target and the first predicted target is above a threshold.

公开(公告)号：US20240020797A1

公开(公告)日：2024-01-18

申请号：US17863501

申请日：2022-07-13

Applicant: GM GLOBAL TECHNOLOGY OPERATIONS LLC

Inventor： Tzvi Philipp ,  Dan Levi ,  Shai Silberstein

IPC: G06T5/00

CPC classification number: G06T5/003 ,  G06T2207/20081 ,  G06T2207/20201

Abstract: A method for deblurring a blurred image includes dividing the blurred image into overlapping regions each having a size and an offset from neighboring overlapping regions along a first direction as determined by a period of a ringing artifact in the blurred image, or by obtained blur characteristics relating to the blurred image and/or attributable to the optical system, or by a detected cause capable of producing the blur characteristics, stacking the overlapping regions to produce a stacked output, wherein the overlapping regions are sequentially organized along the first direction, convolving the stacked output through a first convolutional neural network (CNN) to produce a first CNN output having reduced blur as compared to the stacked output, and assembling the first CNN output into a re-assembled image, and processing the re-assembled image through a second CNN to produce a deblurred image having reduced residual artifacts as compared to the re-assembled image.