Discriminating Tumor from PET Scan Images Using Multi-level Convolutional Neural Network

홈 > 연구문헌 > 학술대회 프로시딩 > 한국정보과학회 학술대회 > 2019년 컴퓨터종합학술대회

한글제목(Korean Title)	Discriminating Tumor from PET Scan Images Using Multi-level Convolutional Neural Network
영문제목(English Title)	Discriminating Tumor from PET Scan Images Using Multi-level Convolutional Neural Network
저자(Author)	Tien X. Dang Thong-Van Huynh Soo-Hyung Kim Jung-Joon Min Sae-Ryung Kang
원문수록처(Citation)	VOL 46 NO. 01 PP. 0805 ~ 0807 (2019. 06)
한글내용 (Korean Abstract)
영문내용 (English Abstract)	Nowadays, many breakthroughs have been made in image classification, however, medical imaging classification still a challenge such as classification tumor and non-tumor images. Classification is the first step and it plays an important role in tumor segmentation task. In this paper, we propose a deep learning model for discriminating tumor and non-tumor in Positron Emission Tomography (PET) scan images. Particularly, we propose a new deep learning model called multi-level in convolutional neural networks (ML-Net), which is a combination of a hierarchy of features intended to improve the classification task. The proposed algorithm was evaluated on the PET dataset and received good accuracy.
키워드(Keyword)
파일첨부	PDF 다운로드