COVID-19 is a respiratory disease caused by severe acute respiratory syndrome coronavirus (SARS-CoV-2). This paper proposes a deep learning model to assist medical imaging physicians in diagnosing COVID-19 cases. We designed the Parallel Channel Attention Feature Fusion Module (PCAF), and brand new structure of convolutional neural network MCFF-Net was put forward. The experimental results show that the overall accuracy of MCFF-Net66-Convl-GAP model is 96.79% for 3-class classification. Simultaneously, the precision, recall, specificity and the sensitivity for COVID-19 are both 100%. Compared ...