公开(公告)号：US20220147822A1

公开(公告)日：2022-05-12

申请号：US17459066

申请日：2021-08-27

Applicant: BEIJING BAIDU NETCOM SCIENCE AND TECHNOLOGY CO., LTD.

Inventor： Ying XIN ,  Yuan FENG ,  Guanzhong WANG ,  Pengcheng YUAN ,  Bin ZHANG ,  Xiaodi WANG ,  Xiang LONG ,  Yan PENG ,  Honghui ZHENG ,  Shumin HAN

IPC: G06N3/08 ,  G06N3/04 ,  G06V10/82 ,  G06V10/766 ,  G06V10/77

Abstract: Provided are a training method and apparatus for a target detection model, a device and a storage medium. The training method is described below. A feature map of a sample image is processed through a classification network of an initial model and a heat map and a classification prediction result of the feature map are obtained, a classification loss value is determined according to the classification prediction result and classification supervision data of the sample image, and a category probability of pixels in the feature map is determined according to the heat map of the feature map and a probability distribution map of the feature map is obtained; the feature map is processed through a regression network of the initial model and a regression prediction result is obtained, and a regression loss value is determined.

公开(公告)号：US20210312240A1

公开(公告)日：2021-10-07

申请号：US17348285

申请日：2021-06-15

Applicant: Beijing Baidu Netcom Science and Technology Co., Ltd.

Inventor： Xiaodi WANG ,  Shumin HAN ,  Yuan FENG ,  Ying XIN ,  Bin ZHANG ,  Shufei LIN ,  Pengcheng YUAN ,  Xiang LONG ,  Yan PENG ,  Honghui ZHENG

IPC: G06K9/62 ,  G06K9/46 ,  G06N3/02

Abstract: A header model for instance segmentation includes a target box branch having a first branch and a second branch, where the first branch is configured to process an inputted first feature map to obtain class information and confidence of a target box, and the second branch is configured to process the first feature map to obtain location information of the target box. The header model also includes a mask branch configured to process an inputted second feature map to obtain mask information, wherein the second feature map is a feature map outputted by an ROI extraction module, and the first feature map is a feature map resulting from a pooling performed on the second feature map.

公开(公告)号：US20220020175A1

公开(公告)日：2022-01-20

申请号：US17489991

申请日：2021-09-30

Applicant: Beijing Baidu Netcom Science and Technology Co., Ltd.

Inventor： Xiaodi WANG ,  Shumin HAN ,  Yuan FENG ,  Ying XIN ,  Bin ZHANG ,  Xiang LONG ,  Honghui ZHENG ,  Yan PENG ,  Zhuang JIA

IPC: G06T7/73 ,  G06K9/62

Abstract: An object detection model training method, object detection method and related apparatus, relate to the field of artificial intelligence technologies such as computer vision, deep learning. An implementation includes: obtaining training sample data including a first remote sensing image and position annotation information of an anchor box of a subject to be detected in the first remote sensing image, where the position annotation information includes angle information of the anchor box relative to a preset direction; obtaining an object feature map of the first remote sensing image based on an object detection model, performing object detection on the subject to be detected based on the object feature map to obtain an object bounding box, and determining loss information between the anchor box and the object bounding box based on the angle information; updating a parameter of the object detection model based on the loss information.