跳到主要导航 跳到搜索 跳到主要内容

A multi-stage fusion strategy for multi-scale GLCM-CNN model in differentiating malignant from benign polyps

  • Jiaxing Tan
  • , Shu Zhang
  • , Weiguo Cao
  • , Yongfeng Gao
  • , Lihong Li
  • , Yumei Huo
  • , Zhengrong Liang
  • Stony Brook University
  • City University of New York

科研成果: 书/报告/会议事项章节 › 会议稿件 › 同行评审

4 引用 (Scopus)

摘要

Computer aided diagnosis (CADx) of polyps has shown great potential to advance the computed tomography colonography (CTC) technique with diagnostic capability. Facing the problem of numerous uncertainties such as polyp size, shape, and orientation in CTC, GLCM-CNN has been proved to be an effective deep learning based tumor classification method, where convolution neural network (CNN) makes decision based on the texture pattern encoded in gray level co-occurrence matrix (GLCM) containing 13 directions. The 13 directional GLCM, by sampling displacement, can be classified into 3 subgroups. Based on our evaluation on the information encoded in the three subgroups, we propose a multi-stage fusion CNN model, which makes the final decision based on two types of features, i.e. (1) a gate module selected group-specific features and (2) fused features learnt from all the features from three groups. On our polyp dataset, which contains 87 polyp masses, our proposed method outperforms both single sub-group based and 13 directional GLCM based CNN model by at least 1.3% in AUC by the average of 20 times 2 fold cross validation experiment results.

源语言英语
主期刊名Medical Imaging 2020
主期刊副标题Computer-Aided Diagnosis
编辑Horst K. Hahn, Maciej A. Mazurowski
出版商SPIE
ISBN(电子版)9781510633957
DOI
出版状态已出版 - 2020
已对外发布是
活动Medical Imaging 2020: Computer-Aided Diagnosis - Houston, 美国
期限: 16 2月 2020 → 19 2月 2020

丛书

姓名Progress in Biomedical Optics and Imaging - Proceedings of SPIE
卷11314
ISSN(印刷版)1605-7422

会议

会议Medical Imaging 2020: Computer-Aided Diagnosis
国家/地区美国
市Houston
时期16/02/20 → 19/02/20

学术指纹

探究 'A multi-stage fusion strategy for multi-scale GLCM-CNN model in differentiating malignant from benign polyps' 的科研主题。它们共同构成独一无二的学术指纹。

引用此