ETRI-Knowledge Sharing Plaform

ENGLISH

성과물

논문 검색
구분 SCI
연도 ~ 키워드

상세정보

학술대회 A Comparative Study on the Maritime Object Detection Performance of Deep Learning Models
Cited 1 time in scopus Download 1 time Share share facebook twitter linkedin kakaostory
저자
문성원, 이지원, 이정수, 남도원, 유원영
발행일
202010
출처
International Conference on Information and Communication Technology Convergence (ICTC) 2020, pp.1155-1157
DOI
https://dx.doi.org/10.1109/ICTC49870.2020.9289620
협약과제
20HH5100, 영상 내 객체간 관계 분석 기반 해상 선박/구조물 상세 식별 콘텐츠 기술 개발, 남도원
초록
With the increasing volume of maritime traffic, the need for maritime surveillance is also increasing. In this situation, unlike the land environment where various data sets are built, the marine environment lacks data, so the progress of technology research is insufficient. Several object detection methods have been proposed and show excellent performance in areas where sufficient data is secured, but performance verification of existing techniques has not been performed on marine images. In this paper, we compare the performance of marine object detection in various marine images with the latest object detection methods and propose an object detection method suitable for marine environments.
키워드
CNN, maritime object, object detection
KSP 제안 키워드
Data sets, Detection Method, Marine environment, Maritime Traffic, Maritime surveillance, Object detection, Performance verification, comparative study, deep learning(DL), deep learning models, detection performance