Multi-crop Convolutional Neural Networks for lung nodule malignancy suspiciousness classification

作者：

Highlights：

• A novel MC-CNN model is proposed to classify lung nodule malignancy suspiciousness.

• Our MC-CNN extracts multi-scale nodule salient features within a single network.

• The proposed method does not rely on nodule segmentation or hand-crafted features.

• Performance is evaluated and validated on LIDC-IDRI dataset.

• Potential nodule relevant applications of the proposed method are provided.

摘要

Highlights•A novel MC-CNN model is proposed to classify lung nodule malignancy suspiciousness.•Our MC-CNN extracts multi-scale nodule salient features within a single network.•The proposed method does not rely on nodule segmentation or hand-crafted features.•Performance is evaluated and validated on LIDC-IDRI dataset.•Potential nodule relevant applications of the proposed method are provided.

论文关键词：Lung nodule,Malignancy suspiciousness,Convolutional neural network,Multi-crop pooling

论文评审过程：Received 28 January 2016, Revised 29 April 2016, Accepted 24 May 2016, Available online 26 May 2016, Version of Record 13 October 2016.

论文官网地址：https://doi.org/10.1016/j.patcog.2016.05.029