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
Dongrae Cho (Deepmedi) Hyun-Du Jeong (JASAN) Jae Ho Jung (JASAN) Kwang Jin Lee (Deepmedi) Jongin Kim (Deepmedi)
저널정보
한국정보통신학회 INTERNATIONAL CONFERENCE ON FUTURE INFORMATION & COMMUNICATION ENGINEERING 2024 INTERNATIONAL CONFERENCE ON FUTURE INFORMATION & COMMUNICATION ENGINEERING Vo.15 No.1
발행연도
2024.1
수록면
79 - 82 (4page)

이용수

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

초록· 키워드

오류제보하기
In this study, we developed a U-net based deep learning model for detecting pores distributed on the face. Generally, a widely used deep learning model for image segmentation is the U-net model, which is typically constructed based on convolution layers. To minimize information loss, a key feature of this architecture is the reattachment of previously trained layers to the subsequent layers, making it an optimal choice for image segmentation. However, the model proposed in this study is configured using besidual blocks instead of convolution layers. Convolution layers face the challenge of the gradient vanishing problem as layers become deeper, where previously learned weight values are lost. In contrast, residual blocks are a special layer type designed to preserve previously learned weights by passing them tot eh subsequent layers. To train and evaluate the proposed deep learning model, we used the JANUS-Pro device to simultaneously acquire facial images and images of facial pores. Additionally, we employed face detection and face mesh detection algorithms to extract facial images, which were then utilized as training and evaluation data. As a result, our proposed U-net (U-net based on residual block) exhibited faster convergence of loss values compared to the traditional U-net, and it demonstrated more accurate pore detection. In the future, the U-net proposed in this study will be utilized to capture facial images using smartphones and infer the locations of pores, enabling a simple quantification of pore distribution and size, which can only be examined on a smartphone.

목차

Abstract
Ⅰ. INTRODUCTION
Ⅱ. Realted work
Ⅲ. METHODS
Ⅳ. RESULT AND 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