In order to establish the correct protocol for COVID-19 treatment, estimating the percentage of COVID-19 specific infection within the lung tissue can be an important tool. This article describes the approach we used in order to approximate the COVID-19 percentage on lung CT scan slices from the Covid-19-Infection-Percentage-Estimation-Challenge. Our method is based on modern training pipelines and architectures, used for training state of the art models on image classification task. We obtained the best score on the validation dataset for the Covid-19 Infection Percentage Estimation competition. The final model consists of an ensemble of 40 models and achieved a score of 4.17140661 MAE, 0.948787386 PC and 8.196144648 RMSE.
La acest articol au contribuit si cercetatori de la Synbrain si colaboratori. CITESTE INTREGUL ARTICOL