ETRI-Knowledge Sharing Plaform

KOREAN
논문 검색
Type SCI
Year ~ Keyword

Detail

Conference Paper 국방 경계감시 능력 증강을 위한 AI 허브 데이터셋 활용 연구
Cited - time in scopus Share share facebook twitter linkedin kakaostory
Authors
이지원, 문성원, 남도원, 유원영
Issue Date
2023-06
Citation
대한전자공학회 학술 대회 (하계) 2023, pp.1140-1143
Publisher
대한전자공학회
Language
Korean
Type
Conference Paper
Abstract
As the number of cases in which deep learning is applied to real life increases, there is a movement to utilize deep learning technology in the defense surveillance field as well. In order to apply deep learning technology, a certain level of learning data is required, and learning data are being collected for this purpose. However, the number of learning data for targets with low vigilant surveillance appearance, such as aircraft and drones, is insufficient. In this paper, we would like to discuss how to utilize AI hub data to solve this problem. Through experiments, it was confirmed that performance improvement may not occur even if unrefined training data is added. In the future, we plan to study ways to collect additional data that is meaningful for deep model learning.
KSP Keywords
Deep model, Learning Technology, Learning data, Model learning, To discuss, deep learning(DL), performance improvement, training data