Kaggle: COVID-19 Detection
In this competition, you’ll identify and localize COVID-19 abnormalities on chest radiographs. In particular, you'll categorize the radiographs as negative for pneumonia or typical, indeterminate, or atypical for COVID-19. Organizers provided dataset - imaging data and annotations from a group of radiologists.
In other words the task to solve is image classification accompanied by attention
- user shall classify patient COVID19 situation and also tell why think so, what are the regions in the scan that makes him think this case is positive.
A simple way how to use this basic functions:
! pip install https://github.com/Borda/kaggle_COVID-detection/archive/main.zipTraining progress with ResNet50 with training for 50 epochs:
