ABSTRACT
The entire world is suffering from the corona pandemic (COVID-19) since December 2019. Deep convolutional neural networks (deep CNN) can be used to develop a rapid detection system of COVID-19. Among all the existing literature, ResNet50 is showing better performance, but with three main limitations, i.e.: 1) overfitting;2) computation cost;3) loss of feature information. To overcome these problems authors have proposed four different modifications on ResNet50, naming it as LightWeightResNet50. An image dataset containing chest X-ray images of coronavirus patients and normal persons is used for evaluation. Five-fold cross-validation is applied with transfer learning. Ten different performance measures (true positive, false negative, false positive, true negative, accuracy, recall, specificity, precision, F1-score and area under curve) are used for evaluation along with fold-wise performance measures comparison. The four proposed methods have an accuracy improvement of 4%, 13%, 14% and 7% respectively when compared with ResNet50.
ABSTRACT
The entire world is suffering from the corona pandemic (COVID-19) since December 2019. Deep convolutional neural networks (deep CNN) can be used to develop a rapid detection system of COVID-19. Among all the existing literature, ResNet50 is showing better performance, but with three main limitations, i.e.: 1) overfitting;2) computation cost;3) loss of feature information. To overcome these problems authors have proposed four different modifications on ResNet50, naming it as LightWeightResNet50. An image dataset containing chest X-ray images of coronavirus patients and normal persons is used for evaluation. Five-fold cross-validation is applied with transfer learning. Ten different performance measures (true positive, false negative, false positive, true negative, accuracy, recall, specificity, precision, F1-score and area under curve) are used for evaluation along with fold-wise performance measures comparison. The four proposed methods have an accuracy improvement of 4%, 13%, 14% and 7% respectively when compared with ResNet50.
ABSTRACT
Human faces being highly dynamic, are extensively studied in the field of pattern recognition, computer vision and artificial intelligence. Moreover, identification of faces using a part of it still remains an understudied domain. Detection of faces using just uncovered eye images can be a boon for surveillance and security especially in times of Covid-19 when most people are advised to cover their faces in a pub-lic space. In this paper we present a system, which identifies the person's face using the visible eye region namely the eyes and the forehead portions of the per-son. The model is trained over basic convolution net-work and the classification is done using Siamese net-works. The classification accuracy is measured using the dis-similarity score which calculates the euclidean distance between the converted feature vectors of the eye regions. The regions which are similar have neg-ligible dissimilarity score. © 2022 IEEE.