Skip to content

Latest commit

 

History

History
20 lines (13 loc) · 838 Bytes

README.md

File metadata and controls

20 lines (13 loc) · 838 Bytes

ECCV22-COVID19

Implementation of "Two Stage COVID19 Classification Using BERT Features" for ECCV-2022 MIA COV19D Competition.

There are Four parts in this project

Preprocess

Preprocess the CT-scan volume images: check the image size, extract bounding box and percentage of the the lung in the whole image, select images for 3D CNN

Segmentation

A UNet segmentation network is trained. It is used to segment lung mask of an image.

BERT

A 3D CNN network with BERT for CT-scan volume classification and embedding feature extraction

embedding

Generate embeddings in the first stage 3D-CNN-BERT network. These embeddings are used as input in the 2nd stage BERT classification.

License

The code of 3D-CNN-BERT-COVID19 is released under the MIT License. There is no limitation for both academic and commercial usage.