-
公开(公告)号：US20240202497A1
公开(公告)日：2024-06-20
申请号：US18572377
申请日：2021-07-21
Applicant: ROBERT BOSCH GMBH , TSINGHUA UNIVERSITY
Inventor： Chunjiang Ge , Gao Huang , Rui Lu , Shiji Song , Xuran Pan , Hao Yang
IPC: G06N3/045 , G06N3/0464
CPC classification number: G06N3/045 , G06N3/0464
Abstract: A method for computer vision processing. The method includes projecting input visual data into a plurality of intermediate feature maps by performing a plurality of 1×1 convolution operations; generating an attention weighted map by performing attention and aggregation operations on the plurality of intermediate feature maps; generating a convolved feature map by performing shift and summation operations on the plurality of intermediate feature maps; and adding the attention weighted map and the convolved feature map based on at least one scalar.
-
公开(公告)号：US20240127455A1
公开(公告)日：2024-04-18
申请号：US18546811
申请日：2021-03-03
Applicant: Robert Bosch GmbH , TSINGHUA UNIVERSITY
Inventor： Chufeng Tang , Hang Chen , Jianmin Li , Xiao Li , Xiaolin Hu , Hao Yang
CPC classification number: G06T7/12 , G06T7/13 , G06V20/70 , G06T2207/20021 , G06V20/50
Abstract: Methods and apparatuses of boundary refinement for instance segmentation. The methods for instance segmentation include receiving an image and an instance mask identifying an instance in the image; extracting a set of image patches from the image based on a boundary of the instance mask; generating a refined mask patch for each of the set of image patches based on at least a part of the instance mask corresponding to the each of the set of image patches; and refining the boundary of the instance mask based on the refined mask patch for each of the set of image patches.
-
公开(公告)号：US20240078436A1
公开(公告)日：2024-03-07
申请号：US18259563
申请日：2021-01-04
Applicant: Robert Bosch GmbH , TSINGHUA UNIVERSITY
Inventor： Hang Su , Jun Zhu , Zhengyi Wang , Hao Yang
IPC: G06N3/094
CPC classification number: G06N3/094
Abstract: A method for generating adversarial examples for a Graph Neural Network (GNN) model. The method includes: determining vulnerable features of target nodes in a graph based on querying the GNN model, wherein the graph comprising nodes including the target nodes and edges, each of the edges connecting two of the nodes; grouping the target nodes into a plurality of clusters according to the vulnerable features of the target nodes; and obtaining the adversarial examples based on the plurality of clusters.
-
公开(公告)号：US20240370991A1
公开(公告)日：2024-11-07
申请号：US18290769
申请日：2021-07-30
Applicant: Robert Bosch GmbH
Abstract: A method for automatic optical inspection includes (i) receiving an image of an object, (ii) classifying the image of the object as one of a plurality of categories by a classification model, (iii) determining a label for the image of the object as being qualified if the category obtained by the classification model is a first category, and (iv) performing defect measurement for the image by a segmentation model and determining a label for the image as being qualified or unqualified based on the defect measurement obtained by the segmentation model if the category obtained by the classification model is a second category.
