메뉴 건너뛰기
.. 내서재 .. 알림
소속 기관/학교 인증
인증하면 논문, 학술자료 등을  무료로 열람할 수 있어요.
한국대학교, 누리자동차, 시립도서관 등 나의 기관을 확인해보세요
(국내 대학 90% 이상 구독 중)
로그인 회원가입 고객센터 ENG
주제분류

추천
검색
질문

논문 기본 정보

자료유형
학술저널
저자정보
Jongmin Lee (Chung-Ang University) Seongjin Jeong (Chung-Ang University) Jongwon Choi (Chung-Ang University)
저널정보
중앙대학교 영상콘텐츠융합연구소MINT Moving Image & Technology (MINT) MINT: Moving Image & Technology, Vol.2, No.2
발행연도
2022.5
수록면
14 - 18 (5page)
DOI
10.15323/mint.2022.5.2.2.14

이용수

표지
📌
연구주제
📖
연구배경
🔬
연구방법
🏆
연구결과
AI에게 요청하기
추천
검색
질문

초록· 키워드

오류제보하기
Existing deep learning requires a significant amount of training data to adapt to new testing environments. Furthermore, to compose significant amount of information of the training data, the architecture of deep learning becomes complicated with the necessity of an enlarged model size. To address these challenges, we propose a novel framework to utilize simulated images for real-world applications to reduce the labeling cost for real images. However, because these simulated images have significant discrepancies from images collected in the field, such as resolution, texture, and focal length issues, it is challenging to learn meaningful representations from these simulated images. Unsupervised domain adaptation (UDA) has been employed to overcome this challenge; however, previous UDA methods require a large computational complexity to reduce the gap between the simulated and real images. To address this issue, we integrate the UDA scheme with network compression, and obtain a deep learning model adapted to real images with compressed computation. We perform experiments using the VISDA-2017 benchmark dataset, which validates the effectiveness of the integrated framework in real domains.

목차

Abstract
1. Introduction
2. Related Works
3. Framework
4. Experimental Results
5. Conclusion
References

참고문헌 (0)

참고문헌 신청

함께 읽어보면 좋을 논문

논문 유사도에 따라 DBpia 가 추천하는 논문입니다. 함께 보면 좋을 연관 논문을 확인해보세요!

이 논문의 저자 정보

이 논문과 함께 이용한 논문

최근 본 자료

전체보기

댓글(0)

0

UCI(KEPA) : I410-ECN-0101-2022-688-001637741