Login| Sign Up| Help| Contact|

Patent Searching and Data


Title:
TRAINING METHOD AND TRAINING APPARATUS FOR ROTATING-SHIP TARGET DETECTION MODEL, AND STORAGE MEDIUM
Document Type and Number:
WIPO Patent Application WO/2023/116631
Kind Code:
A1
Abstract:
Disclosed in the present invention are a training method and training apparatus for a rotating-ship target detection model, and a storage medium. The training method comprises: performing preprocessing and angle labeling processing on an original rotating-ship image, so as to obtain a labelled ship image; according to a rotation loss condition after the previous round of training, setting a rotation trigger probability of the current round of training, and according to the labelled ship image, performing data enhancement processing, so as to obtain an enhanced ship image; inputting the enhanced ship image into a rotating-ship target detection model to be trained, so as to obtain a predicted value; according to the predicted value and a real value corresponding to the labelled ship image, calculating a loss function value of the current round of training; and according to the loss function value, updating a model parameter of the rotating-ship target detection model to be trained, so as to complete the current round of training. By means of loss feedback, dynamic data enhancement is performed during a training process, thereby ameliorating the problem of sample imbalance, such that data enhancement becomes targeted, thereby improving the efficiency of data enhancement.

Inventors:
ZHOU YIMIN (CN)
WU XIANGDONG (CN)
Application Number:
PCT/CN2022/140069
Publication Date:
June 29, 2023
Filing Date:
December 19, 2022
Export Citation:
Click for automatic bibliography generation   Help
Assignee:
SHENZHEN INST OF ADV TECH CAS (CN)
International Classes:
G06V20/17; G06N3/04; G06N3/08; G06V10/44; G06V10/82
Foreign References:
CN114419467A2022-04-29
CN113468968A2021-10-01
CN113536936A2021-10-22
US20200143205A12020-05-07
Attorney, Agent or Firm:
BEIJING ZHONG XUN TONG DA INTELLECTUAL PROPERTY AGENCY CO., LTD. (CN)
Download PDF: