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
학술대회자료
저자정보
Hyeonseok Kim (Seoul National University of Science and Technology) Yeejin Lee (Seoul National University of Science and Technology)
저널정보
제어로봇시스템학회 제어로봇시스템학회 국제학술대회 논문집 ICCAS 2023
발행연도
2023.10
수록면
842 - 846 (5page)

이용수

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

초록· 키워드

오류제보하기
The prediction of image angles is a crucial area of study for industrial automation. Despite the various studies conducted in computer vision, predicting image angles has received limited attention. One of the reasons for this lack of focus is that not all objects have clear directionality. For example, an image of a car wheel may have an ambiguous angle, making it infeasible to accurately predict the rotation angle. Therefore, objects with ambiguous angles can introduce noise during the training and testing of rotation angle prediction models. To tackle this issue, we propose a class-agnostic self-supervised angle prediction learning framework that filters out images containing objects with ambiguous angles based on feature similarity. This approach involves two networks: the directionality categorization network, which identifies and eliminates images of undirectional objects, and the rotation categorization network, which learns from the filtered inputs to improve the accuracy of angle predictions. The experimental results using the STL-10 and CIFAR-100 datasets demonstrate that the proposed framework improves rotation angle classification accuracy without the need for rotation angle labels, which are often difficult to obtain in the literature.

목차

Abstract
1. INTRODUCTION
2. PROPOSED METHOD
3. EXPERIMENTAL RESULTS
4. 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

최근 본 자료

전체보기

댓글(0)

0

UCI(KEPA) : I410-151-24-02-088265001