-from-Computerized-Tomography-using-Deep-Learning
所属分类：模式识别(视觉/语音等)
开发工具：Jupyter Notebook
文件大小：5950KB
下载次数：0
上传日期：2022-07-07 05:22:48
上 传 者：
sh-1993
说明： COVID-19的风险分析-来自计算机断层扫描-使用深度学习，。。。算法的性能最好，准确率为97.3%...
(This study aimed to describe the prediction of COVID-19 using chest CT images by using GAN for resampling depending on deep transfer learning models and machine learning (ML) algorithms and risk analysis of patients based on traditional machine learning. The dataset consisted of 8129 samples in two classes, including 5927 CT chest images of)
文件列表:
4-2project (0, 2022-07-07)
4-2project\Completesegmentation.ipynb (1025989, 2022-07-07)
4-2project\Completesegmentation1.ipynb (256239, 2022-07-07)
4-2project\Completesegmentation2.ipynb (638524, 2022-07-07)
4-2project\Transfer Learning.ipynb (98816, 2022-07-07)
4-2project\binarysegmentation.ipynb (272120, 2022-07-07)
4-2project\binarysegmentation1.ipynb (426433, 2022-07-07)
4-2project\classification_without resampling.ipynb (99623, 2022-07-07)
4-2project\colorsegmentation.ipynb (2247512, 2022-07-07)
4-2project\colorsegmentation1.ipynb (2182873, 2022-07-07)
4-2project\dataanalysis.ipynb (104402, 2022-07-07)
4-2project\datareading.ipynb (600097, 2022-07-07)
4-2project\gan.ipynb (412514, 2022-07-07)
4-2project\prediction.ipynb (94840, 2022-07-07)
4-2project\resampled images.ipynb (1042053, 2022-07-07)
4-2project\segmentation.ipynb (402286, 2022-07-07)
# Risk-analysis-of-COVID-19-from-Computerized-Tomography-using-Deep-Learning
This study aimed to describe the prediction of COVID-19 using chest CT images by using GAN for resampling depending on deep transfer learning models and machine learning (ML) algorithms and risk analysis of patients based on traditional machine learning. The dataset consisted of 8129 samples in two classes, including 5927 CT chest images of positive patients with confirmed COVID-19 and 2202 images of negative cases was assessed. There is an imbalance in the dataset. To overcome the imbalance, Generative Adversarial Network (GAN) is introduced. So that the dataset is balanced with 5927 chest CT images in each class. The VGG16 model has obtained the training with an accuracy of 93%. In combining pre-trained models with ML algorithms, the VGG16 model and SVM algorithm have received the best performance with an accuracy of 97.3%. The traditional machine learning method is used for risk analysis of patients based on the level of infection. Binary Segmentation and Semantic Segmentation are applied to identify tumours and pneumonia.
近期下载者：
相关文件：
收藏者：