Invention Grant
- Patent Title: Method and system for training image classification model
-
Application No.: US17191715Application Date: 2021-03-04
-
Publication No.: US11748450B2Publication Date: 2023-09-05
- Inventor: Hye Rim Bae , Hye Mee Kim
- Applicant: Pusan National University Industry-University Cooperation Foundation
- Applicant Address: KR Busan
- Assignee: PUSAN NATIONAL UNIVERSITY INDUSTRY-UNIVERSITY COOPERATION FOUNDATION
- Current Assignee: PUSAN NATIONAL UNIVERSITY INDUSTRY-UNIVERSITY COOPERATION FOUNDATION
- Current Assignee Address: KR Busan
- Agency: CANTOR COLBURN LLP
- Priority: KR 20200150359 2020.11.11
- Main IPC: G06N3/08
- IPC: G06N3/08 ; G06F18/21 ; G06F18/23213

Abstract:
A method and system for training an image classification model is disclosed. An aspect is to separate training processes of a feature value extraction model and an image classification model and train the feature value extraction model on a representative feature value suitable for image classification into a specific label value (e.g., “Peak”), thereby improving accuracy and performance of a classification model for a ground-penetrating radar (GPR) image that is captured by a GPR and is not easy for feature value extraction.
Public/Granted literature
- US20220147771A1 METHOD AND SYSTEM FOR TRAINING IMAGE CLASSIFICATION MODEL Public/Granted day:2022-05-12
Information query