A 3D CNN Network with BERT For Automatic COVID-19 Diagnosis From CT-Scan Images
Weijun Tan, Jingfeng Liu
Abstract
We present an automatic COVID1-19 diagnosis framework from lung CT-scan slice images. In this framework, the slice images of a CT-scan volume are first preprocessed using segmentation techniques to filter out images of closed lung, and to remove the useless background. Then a re-sampling method is used to select a set of fixed number of slice images for training and validation. A 3D CNN network with BERT is used to classify this set of selected slice images. In this network, an embedding feature is also extracted. In cases where there are more than one set of slice images in a volume, the features of all sets are extracted and pooled into a feature vector for the whole CT-scan volume. A simple multiple-layer perceptron (MLP) network is used to further classify the aggregated feature vector. The models are trained and evaluated on the provided training and validation datasets. On the validation dataset, the precision is 0.9278 and the F1 score is 0.9261. On the test dataset, the F1 score is 0.8822. The code will be available at xxx.