Multi-Branch Ensemble Learning Architecture Based on 3D CNN for False Positive Reduction in Lung Nodule Detection
| Autor: | Guangzhi Ma, Xiangyang Xu, Haichao Cao, Chih-Cheng Hung, Enmin Song, Hong Liu, Tengying Liu, Renchao Jin |
|---|---|
| Rok vydání: | 2019 |
| Předmět: |
General Computer Science
Generalization Computer science offline hard mining Convolutional neural network 030218 nuclear medicine & medical imaging Reduction (complexity) 03 medical and health sciences 0302 clinical medicine General Materials Science Spatial analysis Network architecture business.industry General Engineering Pattern recognition Computer-aided diagnosis Ensemble learning 3D CNN ensemble learning lcsh:Electrical engineering. Electronics. Nuclear engineering Artificial intelligence False positive rate business lcsh:TK1-9971 false positive reduction 030217 neurology & neurosurgery |
| Zdroj: | IEEE Access, Vol 7, Pp 67380-67391 (2019) |
| ISSN: | 2169-3536 |
| DOI: | 10.1109/access.2019.2906116 |
| Popis: | It is critical to have accurate detection of lung nodules in CT images for the early diagnosis of lung cancer. In order to achieve this, it is necessary to reduce the false positive rate of detection. Due to the heterogeneity of lung nodules and their similarity to the background, it is difficult to distinguish true lung nodules from numerous candidate nodules. In this paper, in order to solve this challenging problem, we propose a Multi-Branch Ensemble Learning architecture based on the three-dimensional (3D) convolutional neural networks (MBEL-3D-CNN). The method combines three key ideas: 1) constructing a 3D-CNN to make the maximum utilization of spatial information of lung nodules in the 3D space; 2) embedding a multi-branch network architecture in the 3D-CNN that is well adapted to the heterogeneity of lung nodules, and; 3) using ensemble learning to effectively improve the generalization performance of the 3D-CNN model. In addition, we use offline hard mining operations to make the network capable of handling those indistinguishable positive and negative samples. The proposed method was tested on the dataset LUNA16 in our experiments. The experimental results show that MBEL-3D-CNN architecture can achieve better screening results. |
| Databáze: | OpenAIRE |
| Externí odkaz: |
|