RETRACTED: PSCNN: PatchShuffle Convolutional Neural Network for COVID-19 Explainable Diagnosis

Abstract

Objective: COVID-19 is a sort of infectious disease caused by a new strain of coronavirus. This study aims to develop a more accurate COVID-19 diagnosis system.Methods: First, the n-conv module (nCM) is introduced. Then we built a 12-layer convolutional neural network (12l-CNN) as the backbone network. Afterwards, PatchShuffle was introduced to integrate with 12l-CNN as a regularization term of the loss function. Our model was named PSCNN. Moreover, multiple-way data augmentation and Grad-CAM are employed to avoid overfitting and locating lung lesions.Results: The mean and standard variation values of the seven measures of our model were 95.28 ± 1.03 (sensitivity), 95.78 ± 0.87 (specificity), 95.76 ± 0.86 (precision), 95.53 ± 0.83 (accuracy), 95.52 ± 0.83 (F1 score), 91.7 ± 1.65 (MCC), and 95.52 ± 0.83 (FMI).Conclusion: Our PSCNN is better than 10 state-of-the-art models. Further, we validate the optimal hyperparameters in our model and demonstrate the effectiveness of PatchShuffle.

COV-RadNet: A Deep Convolutional Neural Network for Automatic Detection of COVID-19 from Chest X-Rays and CT Scans

“…In future we may include feature selection methods to train the model only with the most informative features. We may use hybrid models (combination of two or more pre-trained models) from pre-trained models and test their performance on the dataset without augmentation and multiple way data augmentation [23] also.…”

“…We have used gamma values from 0.8 to 1.1 to generate new samples. [23] used patchshuffle, a multiple way data augmentation technique. Since chest radiography images contain various kind of lung opacity information and we have four different classes where COVID-19, pneumonia and non COVID lung opacity show very close symptoms, we have implemented only two types of augmentation techniques.…”

A Deep Learning-Based Quantitative Structure–Activity Relationship System Construct Prediction Model of Agonist and Antagonist with High Performance

“…However, since the image will be similar to the original image, the risk of overfitting, i.e., a decrease in the performance on the test dataset due to the prediction model fitting to match into the training dataset, cannot be ruled [57][58][59][60][61][62][63][64][65][66][67][68]. Thus, data augmentation effectively enables learning with a small number of data.…”

A multilevel self‐attention based segmentation and classification technique using Directional Hexagonal Mixed Pattern algorithm for lung nodule detection in thoracic CT image

“…The experiment is done as 10 runs of 10-fold crossvalidation and it achieves an accuracy of approximately 94.03%. Wang et al 19 introduced a 12 level CNN for the diagnosis of COVID-19. They have integrated Pat-chShuffle along with the 12 levels to regularize the loss function.…”