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
Ziaullah Khan (Inje University) Muhammad Omair Khan (Inje University) Hee-Cheol Kim (Inje University)
저널정보
한국정보통신학회 한국정보통신학회 종합학술대회 논문집 한국정보통신학회 2024년도 춘계종합학술대회 논문집 제28권 제1호
발행연도
2024.5
수록면
229 - 233 (5page)

이용수

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

초록· 키워드

오류제보하기
Adversarial attacks pose a significant threat to the reliability and effectiveness of machine learning models in various domains, including medical imaging. In this study, we examine how susceptible a customized Convolutional Neural Network (CNN) model is to such attacks, emphasizing the Fast Gradient Sign Method (FGSM). The purpose is to assess how well the CNN model can categorize images of lung histology into five different groups. We also assess how it affects the precision of lung histopathology image classification. On the initial validation dataset, the model revealed a high baseline accuracy of 93.75%; however, is also showed significant vulnerability to perturbations that are invisible to the human eye. Upon encountering FGSM attacks with an epsilon value o 0.01, the model’s accuracy declined sharply to 65.50%. The accuracy fell to 43.75 because the performance drop was worsened by raising the epsilon value to 0.02. The significant decline in the accuracy of the model, caused by very little noise, highlights the serious threat that adversarial attacks present to the dependability and resilience of machine learning models used in medical imaging. The results emphasize the importance of including adversarial robustness in the deep learning model construction process to guarantee the models’ reliability in clinical settings where precision is critical.

목차

ABSTRACT
I. Introduction
II. Methodology
III. Results and Discussion
IV. Conclusion
References

참고문헌 (0)

참고문헌 신청

함께 읽어보면 좋을 논문

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

최근 본 자료

전체보기

댓글(0)

0