论著摘要 |【AI-MR】基于MRI的前列腺癌检测与高层次的表示和分层分类(双语版)

2018-02-23 10:53:45 admin 16
标签:   人工智能 MR 深度学习 随机森林 前列腺癌

MRI-based prostate cancer detection with high-level representation and hierarchical classification.

发表日期: 2017.03.01   来源:Med Phys. 2017 Mar;44(3):1028-1039.

作者:

Zhu Y1, Wang L2, Liu M2, Qian C3, Yousuf A4, Oto A4 , Shen D2,5.

作者介绍:

1. Computer Center, Nanjing University of Aeronautics & Astronautics, Jiangsu, China.

2. Department of Radiology and BRIC, University of North Carolina, Chapel Hill, NC, USA.

3. School of Science, Nanjing University of Science and Technology, Jiangsu, China.

4. Department of Radiology, Section of Urology, University of Chicago, Chicago, IL, USA.

5. Department of Brain and Cognitive Engineering, Korea University, Seoul, 02841, Korea.

摘要

Abstact

目的

利用深度神经网络对前列腺癌进行高级特征表示,然后基于高级特征表示构造分级分类,对检测结果进行优化。

Purpose

Extracting the high-level feature representation by using deep neural networks for detection of prostate cancer, and then based on high-level feature representation constructing hierarchical classification to refine the detection results.

方法

高级特征表示首先通过深度学习网络习得,其中多参数MR图像被用作输入数据。然后,基于学习到的高层次特征,开发了一种分层分类方法,迭代构建多个随机森林分类器,对前列腺癌的检测结果进行优化。

Methods

High-level feature representation is first learned by a deep learning network, where multiparametric MR images are used as the input data. Then, based on the learned high-level features, a hierarchical classification method is developed, where multiple random forest classifiers are iteratively constructed to refine the detection results of prostate cancer.

结果

对21例真实患者进行了实验,所提出的方法平均切片评估(SBE)为89.90%,平均敏感度为91.51%,平均特异度为88.47%。

Results

The experiments were carried on 21 real patient subjects, and the proposed method achieves an averaged section-based evaluation (SBE) of 89.90%, an averaged sensitivity of 91.51%, and an averaged specificity of 88.47%.

结论

从我们提出的方法学到的高级特征可以比传统的手工特征(例如,LBP和Haar-like特征)在检测前列腺癌区域方面获得更好的性能,并且从所提出的分层分类方法获得的上下文特征在精炼癌症检测结果方面很有效。

Conclusions

The high-level features learned from our proposed method can achieve better performance than the conventional handcrafted features (e.g., LBP and Haar-like features) in detecting prostate cancer regions, also the context features obtained from the proposed hierarchical classification approach are effective in refining cancer detection result.

关键词:

深度学习;分层分类;磁共振成像;前列腺癌检测;随机森林

Keywords:

deep learning; hierarchical classification; magnetic resonance imaging (MRI); prostate cancer detection; random forest

阅读原文:PMID: 28107548  PMCID: PMC5540150  DOI: 10.1002/mp.12116


慧影医疗科技(北京)有限公司

地点:北京市海淀区中关村东升科技园B2-C103

电话:400-890-9020

邮箱:radcloud@huiyihuiying.com

关闭
图片
图片