公开(公告)号：US12266190B2

公开(公告)日：2025-04-01

申请号：US17884356

申请日：2022-08-09

Applicant: Waymo LLC

Inventor： Albert Zhao ,  Vasiliy Igorevich Karasev ,  Hang Yan ,  Daniel Rudolf Maurer ,  Alper Ayvaci ,  Yu-Han Chen

IPC: G06V20/58 ,  G06T7/55 ,  G06V10/44 ,  G06V10/82 ,  G06T3/4046 ,  G06V10/40 ,  G06V10/70 ,  G06V20/69 ,  G06V30/18

Abstract: The described aspects and implementations enable efficient detection and classification of objects with machine learning models that deploy a bird's-eye view representation and are trained using depth ground truth data. In one implementation, disclosed are system and techniques that include obtaining images, generating, using a first neural network (NN), feature vectors (FVs) and depth distributions pixels of images, wherein the first NN is trained using training images and a depth ground truth data for the training images. The techniques further include obtaining a feature tensor (FT) in view of the FVs and the depth distributions, and processing the obtained FTs, using a second NN, to identify one or more objects depicted in the images.

公开(公告)号：US20240096105A1

公开(公告)日：2024-03-21

申请号：US17884356

申请日：2022-08-09

Applicant: Waymo LLC

Inventor： Albert Zhao ,  Vasiliy Igorevich Karasev ,  Hang Yan ,  Daniel Rudolf Maurer ,  Alper Ayvaci ,  Yu-Han Chen

IPC: G06V20/58 ,  G06T7/55 ,  G06V10/44 ,  G06V10/82

CPC classification number: G06V20/58 ,  G06T7/55 ,  G06V10/44 ,  G06V10/82 ,  G06T2207/20081 ,  G06T2207/20084 ,  G06T2207/30252

Abstract: The described aspects and implementations enable efficient detection and classification of objects with machine learning models that deploy a bird's-eye view representation and are trained using depth ground truth data. In one implementation, disclosed are system and techniques that include obtaining images, generating, using a first neural network (NN), feature vectors (FVs) and depth distributions pixels of images, wherein the first NN is trained using training images and a depth ground truth data for the training images. The techniques further include obtaining a feature tensor (FT) in view of the FVs and the depth distributions, and processing the obtained FTs, using a second NN, to identify one or more objects depicted in the images.