key: cord-0950587-94x3svb8
authors: nan
title: An Interpretable Deep Learning Model for Covid-19 Detection With Chest X-Ray Images
date: 2021-06-08
journal: IEEE Access
DOI: 10.1109/access.2021.3087583
sha: 03747efb89a5b01f9add5054d4652144ba54b593
doc_id: 950587
cord_uid: 94x3svb8

Timely and accurate detection of an epidemic/pandemic is always desired to prevent its spread. For the detection of any disease, there can be more than one approach including deep learning models. However, transparency/interpretability of the reasoning process of a deep learning model related to health science is a necessity. Thus, we introduce an interpretable deep learning model: Gen-ProtoPNet. Gen-ProtoPNet is closely related to two interpretable deep learning models: ProtoPNet and NP-ProtoPNet The latter two models use prototypes of spacial dimension [Formula: see text] and the distance function [Formula: see text]. In our model, we use a generalized version of the distance function [Formula: see text] that enables us to use prototypes of any type of spacial dimensions, that is, square spacial dimensions and rectangular spacial dimensions to classify an input image. The accuracy and precision that our model receives is on par with the best performing non-interpretable deep learning models when we tested the models on the dataset of [Formula: see text]-ray images. Our model attains the highest accuracy of 87.27% on classification of three classes of images, that is close to the accuracy of 88.42% attained by a non-interpretable model on the classification of the given dataset.

The world is still struggling with the pandemic of Covid-19 and its variants, such as: B.1.1.7, B.1.351 and P.1 [18] . Many facets efforts have been made to control and contain the disease. These efforts include detection of the virus. Many models have been proposed to detect Covid-19 from medical images, see [4] , [12] , [22] , [23] , [25] , [36] , [37] , [40] , [58] . These models lack the interpretability of their predictions, but the interpretability of the models related to public health is utmost important. The objective of this work is to find an interpretable method to do image classification so that we can tell why an image is classified in a certain way. In this work, we introduce an interpretable deep learning model: generalized prototypical part network (Gen-ProtoPNet), and experiment it over the dataset of three different classes of X -rays, see Section V. Gen-ProtoPNet is a close variation of ProtoPNet [7] and NP-ProtoPNet [41] . To predict the class of a test image, ProtoPNet calculates the similarity scores between learned prototypical parts (with square spacial dimensions 1 × 1) of images from each class and parts of the test image using L2 distance function. These similarity scores are multiplied with a weight matrix to establish a The associate editor coordinating the review of this manuscript and approving it for publication was Amin Zehtabian . positive connection between prototypes and logits of their correct classes, and a zero connection between prototypes and logits of incorrect classes. However, NP-ProtoPNet establishes a negative connection between prototypes and logits of incorrect classes instead of a zero connection between prototypes and logits of incorrect classes, unlike ProtoPnet. Both ProtoPNet and NP-ProtoPNet use the distance function L2 to calculate similarity scores, and they use prototypes of square spacial dimensions 1 × 1. In this work, we use a generalized version of the distance function L2, see Section V-A. The generalized version of the distance function enables us to use prototypical parts of any type of spacial dimensions, that is, rectangular spacial dimensions as well as square spacial dimensions. In this work, a prototypical part or a prototype will represent a part, a patch or a section of an image. The similarity score between a learned prototypical part and an image is considered better if the sum of squares of the differences between the corresponding pixel values of the prototype and a patch of the image is lesser. We are motivated not to use 1 × 1 spacial dimensions for prototypes due to following two reasons: (i) the model gives lesser accuracy; (ii) higher accuracy with the wrong reasoning process.

First, small spacial dimensions (1 × 1) can lead to lesser accuracy because images of objects from different classes can have small parts similar that can lead to good similarity score between patches of a test image and patches of images from wrong classes that gives rise to wrong classification. For example, most of the X -ray images have some part black as a background, see Figure 2 . Therefore, the use of prototypical parts of spacial dimensions 1 × 1 can give good similarity scores between patches of an input image and patches of images from wrong classes, because all the images have some part black. Another example, images of birds from different sea bird species can share same background on most part of the images. So, the use of prototypes of spacial dimension 1 × 1 can wrongly give high similarity score between patches of a test image and patches of images form wrong classes, because mostly such images have water as a background.

Second, the images of objects from completely different classes do not have even small patches similar, and the classification of the images can be made just on the basis of the background in the images instead of identifying the objects themselves. For example, any patch of the images of birds of a sea specie is not similar to any patch of the images of birds of a jungle specie. Therefore, the use of the prototypical parts of the spacial dimensions 1 × 1 can separate a sea bird image from an image of a jungle bird just on the basis of the background in the images instead of identifying the birds themselves.

On the other hand, using prototypical parts of spacial dimension of the biggest possible size can also reduce the accuracy because then the prototypical part will be an image itself instead of being a part of an image. Therefore, there can be only few images that are very similar to such a prototype whose size is equal to the size of an image, but a part of an image with smaller size can be similar to the parts of many other images. Since a prototype represents a part of an image, the optimum value for the spacial dimensions of a prototype lies between 1 × 1 and the biggest possible dimension (in our case, the biggest possible spacial dimension of prototypes is 7 × 7), that is, the spacial dimension of the output of the convolution layers of the baseline models, see Section V.

The use of prototypical parts of spacial dimension bigger than 1 × 1 gives us better accuracy and precision as compared to prototypical parts of spacial dimensions 1 × 1 that are used in ProtoPNet and NP-ProtoPNet, see Section VIII.

Some techniques have been developed to interpret the convolution neural networks, such as: posthoc interpretability analysis and attention-based interpretability. In posthoc analysis of a deep learning model, one interprets a trained convolution neural network after the predictions with the fitting explanations of the reasoning behind classifications made by the model. Posthoc analysis is made with various techniques, such as: activation maximization [10] , [19] , [27] , [35] , [43] , [48] , [57] , deconvolution [59] , and saliency visualization [39] , [43] , [45] , [46] . However, these posthoc visualization methods fail to explain the actual reasoning process behind the classifications made by the models.

Many models have been developed that build attentionbased interpretability, such as: class activation maps and part-based models. An attention-based interpretability model attempts to highlight the parts of an input image on which the network focuses [11] , [15] , [16] , [21] , [38] , [42] , [47] , [56] , [60] - [63] . Nevertheless, these models have serious drawback of not pointing out the prototypical parts that are similar to patches of the image on which the model focuses.

Li et al. [28] , proposed a deep learning architecture that builds case-based reasoning into a neural network. Then Chen et al. [7] along with the authors of the above paper made a considerable improvement in their model ProtoPNet, whereby the network makes prediction by comparing image patches with the learned prototypes. The authors of this paper introduced a model NP-ProtoPNet [41] that is a close variation of ProtoPNet.

As mentioned in the introduction, many networks have been emerged to classify the X -ray images of Covid-19 patients along with X -ray images of normal people and pneumonia patients, see [4] , [12] , [22] , [23] , [25] , [36] , [37] , [40] , [41] , [58] . A study summarizes some papers on Covid-19, and it points out some problems, such as: lack of reliable and adequate amount of data for deep learning algorithms [5] . Some studies related to Covid-19 and IoT are also worth mentioning [1] - [3] , [6] , [8] , [26] , [30] - [33] , [49] . In this work, we also experimented our model over the same dataset that is used in some of the above Covid-19 related articles.

We trained and evaluated our network on the dataset of frontal chest X -ray images of Covid-19 patients [13] , pneumonia patients and normal persons [24] . The dataset of chest X -ray images from Kaggle database [24] has 3875 and 1341 training images of pneumonia patients and normal persons, respectively. Also, the dataset has 390 and 234 test images of pneumonia patients and normal persons, respectively. The other database [13] has 930 medical images of Covid-19 patients. These medical images include, frontal chest X -ray images, CT-scan images, side X -ray images and few obscure (completely black or white) images. Among the medical images of Covid-19 patients, we selected only 748 frontal chest X -ray images. As compared to the number of chest X -ray images of pneumonia patients and normal persons, the number of chest X -ray images of Covid-19 patients was much lesser. Therefore, to balance the data, a copy of the chest X -ray images of Covid-19 patients was included to form a dataset of 1496 images. The 1496 frontal chest X -ray images of Covid-19 patients were divided into train and test sets of 1248 and 248 images, respectively. All these images form the three classes, labeled as: Covid, Normal and Pneumonia. We resized the images to dimension 224 × 224.

For X -ray images given in figures 1 and 2, ProtoPNet is able to recognize many patches of the test image that looks like VOLUME 9, 2021 FIGURE 1. Similarity between learned prototypical parts and an X -ray image of a Covid-19 patient. parts of images of some class. ProtoPNet decides the class of a test image on the basis of a weighted combination of the similarity scores [7] . Similarity scores between patches of a test image and prototypical parts (with square spacial dimensions 1 × 1) are acquired with maximization of the inverted L2 distances between them. The novelty of our model Gen-ProtoPNet is as follows. 1) Our model uses a generalized version (see Section V-A) of the distance function L2. 2) It uses prototypes of both types of spacial dimensions, that is, it uses prototypes of square spacial dimensions as well as rectangular spacial dimensions. 3) It uses prototypes with spacial dimensions bigger than the square spacial dimensions 1×1, that is, either height or width of prototypes is bigger than 1. The use of generalized distance function and spacial dimensions bigger than 1 × 1 helped our model to improve its performance.

We construct our model over the following models: VGG-16, VGG-19 [44] , ResNet-34, ResNet-152 [17] , DenseNet-121, or DenseNet-161 [20] (initialized with filters pretrained on ImageNet [9] ). We call these models baseline or base models. In the Figure 3 , we see that the model comprises of the convolution layers of any of the above base model that are followed by an additional 1×1 layer (we denote these convolution layers together by c whose parameters are collectively denoted by c conv ) and then these convolution layers are followed by a prototype layer p p and a fully connected layer m with weight matrix wt m and no bias. The prototype layer p p is a generalized convolution layer, [14] , [34] . The activation function ReLU is used for all the convolution layers. Note that, Gen-ProtoPNet has only one additional 1×1 convolution layer unlike ProtoPNet.

Let x be an input image and c(x) is the output of the convolutional layers c. If D is the depth of the output c(x) then the shape of c(x) is D × 7 × 7. For example, when Gen-ProtoPNet is constructed over the convolution layers of VGG-16 then the depth of c(x) is 512. Although, there is one additional 1 × 1 convolution layer, but it is not used to reduce the depth of c(x). Therefore, the depth of each prototype is set equal to the depth of output of the regular convolution layers. We pick 10 prototypical parts for each class of images, and this number is randomly chosen. That is, Gen-ProtoPNet learns 30 prototypes. The set of prototypes is denoted by

r=1 . These prototypical parts should catch enough relevant parts for recognizing a test image. The shape of each prototype is D × h × w, where 1 × 1 < h × w < 7 × 7, that is, both h and w are not simultaneously equal to 1. The spacial dimensions of c(x) are 7 × 7, whereas the spacial dimensions of the prototypes are h × w. Thus, every prototypical part p r will be used to represent some prototypical activation pattern in a patch of c(x). Hence, every prototypical part p r can be considered as a representation of a patch of some image. In Figure 3 , the prototypical part p 1 is similar to a part of the X -ray of a Covid-19 patient. The original/source image of the prototypical part p 1 is also given in the Figure 3 . The patch p 1 is the part of the original image, that is enclosed in a rectangle with green boundaries. Similarly, the patches p 11 and p 30 are parts of the original images given in the same row in the Figure 3 . For the output z = c(x) of a test image x, the r-th prototypical unit p p r in p p calculates (with the generalized version of L2) distances between the prototypical part p r and each patch of z. These distances are inverted into similarity scores which results in an activation map of similarity scores. More the activation value, stronger the presence of a prototype in the image x. This activation map preserves the spatial relation of the convolutional output, and can be upsampled to the size of the input image to produce a heat map that identifies which part of the input image is most similar to the learned prototype [7] . These regions are enclosed in the green rectangles on the source images. The activation map is max-pooled to reduce to a single similarity score, that is, there is only one similarity score for each prototype. In the fully connected layer m, the similarity scores produced with global max-pooling are multiplied with the matrix wt m to get the logits, then these logits give prediction after normalization with softmax.

The (r, s)-th entry wt (r,s) m of the weight matrix wt m connects s-th prototype and the logit of r-th class. The similarity scores (after max-pooling) form a column matrix S, see Section VI. The logits of Covid, Normal and Pneumoina classes are obtained from the multiplication of first, second and third rows of wt m with the matrix S. In Figure 3 , the test image is an X -ray image of a Covid-19 patient. The prototypes p 1 − p 10 , p 11 − p 20 and p 21 − p 30 are prototypes of images from Covid, Normal and Pneumonia classes, respectively. The similarity scores between patches of the input image and patches p 1 , p 11 and p 30 are 0.042843, 0.000106 and 0.000082, respectively. The complete list of similarity scores is provided in the similarity score matrix S, see Section VI. The logits for the classes Covid, Normal and Pneumonia are 0.751309, −0.624092 and −0.620997, respectively.

In this section, we describe the generalization of the distance function L2 (Euclidean distance) using base model VGG-16. Also, we present the mathematical formulation and training steps of our algorithm with the generalized distance function. Gen-ProtoPNet is constructed over the regular convolution layers whose output channels have spacial dimension 7 × 7, see Section V. Let x be an input image. Let z (= c(x)) be of the shape (D, 7, 7) , where D is the depth of c(x). Consider a prototype p of the shape (D, h, w) . Let z ijk and p lmk be (i, j) and (l, m) pixels of kth tensor of z and p, respectively. Let zp be obtained by convolving p over z with stride size equal to 1. Then zp is a tensor of the shape (D, 8 − h, 8 − w) . Therefore, each feature map of zp has (8 − h)(8 − w) pixels. For 0 ≤ i ≤ 7 − h, 0 ≤ j ≤ 7 − w and 0 ≤ k ≤ D − 1; (i, j) pixel (zp) ijk of the kth feature map of zp is given by:

Let z 2 be obtained from the Hadamard multiplication of feature maps of z with themselves. Let Z 2 be obtained from z 2 by convolving (over z 2 ) all 1's filter of the shape of prototypes with stride size equal to 1. Note that, Z 2 is the sum of the patches of z 2 of the shape h × w over all feature maps. Therefore, the shape of Z 2 is (8 − h) × (8 − w), and (i, j) pixel Z 2 ij of Z 2 is given by:

(2) VOLUME 9, 2021

Note that, z has (8 − h)(8 − w) patches of the spacial dimension h×w. Hence, the distance d 2 (Z ij , p) between (i, j) patch Z ij (say) of z and a prototype p is given by:

The equations 1 and 2 give the values of (zp) ijk and Z 2 ij .

If the spacial dimension of a prototype p is 1 × 1 then

||z ijk − p 11k || 2 2 , which is the square of L2 distance between a patch of z and the prototype p, where p 11k p k . Therefore, if the spacial dimensions of p are not equal to 1 × 1 then d 2 is a generalization of the distance function L2. The distance function L2 is used in both ProtoPNet and NP-ProtoPNet to find distances of prototypes (spacial dimension 1 × 1) and the patches of images. The prototypical unit p p calculates:

Alternatively,

The equation 4 tells us that if Z is similar to p then d 2 (Z, p) is smaller. The following three steps are performed to train our algorithm.

At this stage of learning, Gen-ProtoPNet aim to learns important features of the image while salient parts cluster near their respective classes. To attain this aim, Gen-ProtoPNet collectively optimize the parameters c conv and p 1 − p 30 in p p using SGD. Let X = {x 1 . . . x n } be a set of image and Y = {y 1 . . . y n } is a set of corresponding labels, and D = {(x r , y r ) : x r ∈ X , y r ∈ Y }. Our goal is to solve the following optimization problem: The decrease in cluster cost clusters prototypical parts around their correct class, see equation 6, whereas the decrease in separation cost attempts to separate prototypical parts from their incorrect class [7] , see equation 7. The decrease in the cross entropy gives better classifications, see equation 5. The coefficients λ 1 is set equal to 0.8 and the coefficient λ 2 belongs to the interval (0.08, 0.8). Let P r be the set of prototypical parts of the images that belong to r-th class. For a class r, we put wt (r,s) m = 1 for all s with p s ∈ P r and wt (r,s) m = −0.5 for all s with p s ∈ P r . Since similarity scores are nonnegative, in this way Gen-ProtoPNet learns a meaningful latent space [7] .

To see which part of the training images are used as prototypes, Gen-ProtoPNet projects every prototype p s onto the patch of the output c(x) that has smallest distance from p s , and x belong to class of p s [7] . That is, for every prototype p s of class r, Gen-ProtoPNet perform the following update:

To rely only on positive connections between prototypes and logits. We aim to make negative connection wt (r,s) m to 0 for all s with p s ∈ P r . We perform this process after fixing all the parameters of convolution layers and prototype layer, and aim to optimize [7] : 

Suppose x is the source image of a prototype p r . The patch of x that is most activated by the prototype p r is used for the visualization of p r . Its activation value must be at least 92nd percentile of all the activation values (before max-pooling) of p p r [7] .

We constructed our model over six baseline models. We trained and tested our model for 500 epochs. The model VGG-16 is used as a baseline model to run the experiments explained in this example. However, the measures of the performance of the model with the other baseline model are given in the Table 1 .

In the Figure 4 , the test image in the first column is a member of the Covid class. In next column, each image is the test image with a rectangle (at a certain place) on it. The rectangles have green boundaries. The pixels enclosed by such a rectangle on an image in the second column correspond to the pixels on the original image in the fourth column and same row. In fact the patch enclosed by a rectangle on an image in the fourth column is the patch from where a prototype (third column and same row) is projected. The fifth column consists of similarity scores that are explained in Section V. The similarity scores of a prototype p s is the s-th entry of the similarity score matrix S (say). The sixth column consists of weights VOLUME 9, 2021 of Covid class. Since Covid is a first class these weight are entries of first row of wt m . The multiplication of the first row of the weight matrix wt m with S gives logit for the Covid class. Similarly, logit for Normal and Pneumonia classes can be obtained by multiplying second and third row of the weight matrix wt m with the matrix S, respectively. Hence, the logits for the first, second and third classes are 0.752591, −0.627040 and −0.623544, respectively. The matrix S and transpose of wt m that we obtain from our experiments are: 

The comparison of the performance of Gen-ProtoPNet with NP-ProtoPNet, ProtoPNet and the base models is made with some metrics, such as: accuracy, precision, recall and F1-score. The confusion matrices are also used to outline the performance of Gen-ProtoPNet. A confusion matrix is an array that is used to describe the performance of a classification model on a set of test data for which the true values are known [54] .

True positive (TP) is the number of items correctly labeled as belonging to the positive class, that is, the items are predicted to belong to a class when they actually belong to that class. True negative (TN ) is the number of items for which the model correctly predict the negative classes, that is, the items are predicted to not belonging to a class when they actually belong to other classes, see [55] . Note that, in non-binary classifications, TP and TN are the diagonal entries of the confusion matrix. False positive (FP) is the number of items incorrectly predicted as belonging to the positive class. False negative (FN ) is the number of items incorrectly predicted as not belonging to the positive class, see [52] . The metrics accuracy, precision and recall in terms of the above positives and negatives are: [2] [0] respectively. Next, we describe the confusion matrix (given in Figure 5 ) for Gen-ProtoPNet when constructed over baseline VGG-16. Total correct predictions made by Gen-ProtoPNet with baseline VGG-16 are 749 (=242 + 119 + 388), see Figure 5 . Total number of test images are 872, see Section V. Thus, the accuracy is 85.89%. The above definitions and Figure 5 give us the precision, recall and F1-score equal to 0.99, 0.98 and 0.98, respectively. Similarly, the metrics for Gen-ProtoPNet with the other baselines can be determined from Figures 6-10 .

The convolution layers of several neural networks can be used to build the models Gen-ProtoPNet, NP-ProtoPNet and ProtoPNet. As stated in Section V, we trained and tested Gen-ProtoPNet with the baseline models over the datesets of the X -rays [13] , [24] . Also, NP-ProtoPNet and ProtoPNet were examined over the same datasets and with the same base models. We trained and tested all models that are compared in Table 1 for 500 epochs.

The measures of the performances of the models (Gen-ProtoPNet, NP-ProtoPNet and ProtoPNet with the six base models) in the metrics can be found in Table 1 . Also, the measures of the performance the base models themselves are given in Table 1 . We explain the Table 1 with an account of the performance of each of these models with base model VGG-16. However, the measures of the performance of these models with the other five base models are also given in Table 1 .

In the second column of the Table 1 , the spacial dimensions of prototypes corresponding to each base model are given. The performance of Gen-ProtoPNet is improved over ProtoPNet with all the base models. Also, the performance of Gen-ProtoPNet is better than the performance of NP-ProtoPNet with some baseline models, and in two cases its performance is better than the performance of the baseline models themselves.

In this section, a graphical comparison of the accuracies of Gen-ProtoPNet with the other models is provided over 100 epochs. In the figures 11-16, the curves of colors purple, yellow, blue and brown sketch the accuracies of Gen-ProtoPNet, NP-ProtoPNet, ProtoPNet and the baselines. For example, in Figure 11 , the accuracies of Gen-ProtoPNet, NP-ProtoPNet and ProtoPNet with the base model VGG-16, and the base model VGG-16 itself are depicted.

The performance in accuracy of Gen-ProtoPNet is the highest with the baseline model VGG-16 and second highest for the remaining base models except VGG-19. Therefore, the curve depicting the accuracy for Gen-PrortoPNet is the highest for base model VGG-16 and second highest for the other base models. The accuracies given by these models become stable before 100 epochs. Although, we experimented these models for 500 epochs, but the comparison of the accuracies of these models is outlined in the figures 11-16 only for first 100 epochs to make the shape of the curves more clearer in the beginning as accuracy of each of these models stabilizes before 100 epochs. 

The convex optimization of the last layer of our model takes considerable time during the training of our model. Our experiments show that zero connection between similarity scores and incorrect classes is hard to achieve during the training process. However, this technique of convex optimization of the last layer is adopted from ProtoPNet model that has already been published in one of the top conferences [7] .

Gen-ProtoPNet is closely related to two interpretable deep learning models ProtoPNet and NP-ProtoPNet that calculate the similarity scores between prototypes of spacial dimension 1×1 and patches of an input image by finding the L2 distance between the prototypes and the patches. In our model, we use a generalized version of the distance function L2 that enables us to use prototypes of any type of spacial dimensions, that is, square spacial dimensions and rectangular spacial dimensions. The use of rectangular spacial dimensions of prototypes enabled our model to improve its performance over ProtoPNet model.

Realizing an effective COVID-19 diagnosis system based on machine learning and IOT in smart hospital environment

COVID-DeepNet: Hybrid multimodal deep learning system for improving COVID-19 pneumonia detection in chest X-ray images

COVID-CheXNet: Hybrid deep learning framework for identifying COVID-19 virus in chest X-rays images

COVID-19 deep learning prediction model using publicly available radiologist-adjudicated chest X-ray images as training data: Preliminary findings

Deep learning and medical image processing for coronavirus (COVID-19) pandemic: A survey

Efficient-CovidNet: Deep learning based COVID-19 detection from chest X-ray images

This looks like that: Deep learning for interpretable image recognition

Conf. Neural Inf. Process. Syst. (NIPS)

Early diagnosis of COVID-19-affected patients based on X-ray and computed tomography images using deep learning algorithm

ImageNet: A large-scale hierarchical image database

Visualizing higherlayer features of a deep network

Look closer to see better: Recurrent attention convolutional neural network for fine-grained image recognition

Predicting COVID-19 pneumonia severity on chest X-ray with deep learning

COVID-19 Image Data Collection

Generalizing the convolution operator in convolutional neural networks

Proc. IEEE Int. Conf. Comput. Vis. (ICCV)

Rich feature hierarchies for accurate object detection and semantic segmentation

Deep residual learning for image recognition

How Many New Coronavirus Variants are There?

A practical guide to training restricted Boltzmann machines

Densely connected convolutional networks

Part-stacked CNN for finegrained visual categorization

A deep learning approach to detect COVID-19 coronavirus with X-ray images

Deep learning based detection and analysis of COVID-19 on chest X-ray images

Chest X-Ray (Pneumoina)

Accurate prediction of COVID-19 using chest X-ray images through deep feature learning model with SMOTE and machine learning classifiers

IoT enabled depthwise separable convolution neural network with deep support vector machine for COVID-19 diagnosis and classification

Convolutional deep belief networks for scalable unsupervised learning of hierarchical representations

Deep learning for case-based reasoning through prototypes: A neural network that explains its predictions

Fully convolutional attention networks for fine-grained recognition

Prediction of COVID-19 active cases using exponential and non-linear growth models

A comprehensive investigation of machine learning feature extraction and classification methods for automated diagnosis of COVID-19 based on Xray images

Benchmarking methodology for selection of optimal COVID-19 diagnostic model based on entropy and TOPSIS methods

IoT-cloud based healthcare model for COVID-19 detection: An enhanced k-nearest neighbour classifier based approach

Efficient implementation of a generalized convolutional neural networks based on weighted Euclidean distance

Synthesizing the preferred inputs for neurons in neural networks via deep generator networks,'' in Proc

Automated detection of COVID-19 cases using deep neural networks with X-ray images

An ensemble based machine learning model for diabetic retinopathy classification

Faster R-CNN: Towards real-time object detection with region proposal networks,'' in Proc

Grad-CAM: Visual explanations from deep networks via gradient-based localization

Artificial intelligence-based classification of chest X-ray images into COVID-19 and other infectious diseases

These do not look like those: An interpretable deep learning model for image recognition

Neural activation constellations: Unsupervised part model discovery with convolutional networks

Deep inside convolutional networks: Visualising image classification models and saliency maps

Very deep convolutional networks for large-scale image recognition

Smooth-Grad: Removing noise by adding noise

Axiomatic attribution for deep networks

Selective search for object recognition

Pixel recurrent neural networks

CovidGAN: Data augmentation using auxiliary classifier GAN for improved COVID-19 detection

When a computer program keeps you in jail: How computers are harming criminal justice

Accuracy and Precision

Precision and Reacall

F-Score

Confusion Matrix

Sensitivity and Specificity

The application of two-level attention models in deep convolutional neural network for fine-grained image classification

Understanding neural networks through deep visualization

COVID-19 detection and disease progression visualization: Deep learning on chest X-rays for classification and coarse localization

Visualizing and understanding convolutional networks

Part-based R-CNNs for fine-grained category detection

Learning multi-attention convolutional neural network for fine-grained image recognition

Learning deep features for discriminative localization

Interpretable basis decomposition for visual explanation

After qualifying UGC-Net, in 2009, he served as an Assistant Professor for a period of two years with the Khalsa College, Garhshankar, India. Since 2015, he has been working as a Research Associate with the Department of Computer Science, and the Department of Mathematics and Statistics. He is currently working as a Postdoctoral researcher with the Kin-Choong Yow's laboratory of Artificial Intelligence Research

He is a member of APEGS and ACM. He has served as a Reviewer for a number of premier journals and conferences, including the IEEE WIRELESS COMMUNICATIONS and the IEEE TRANSACTIONS ON EDUCATION. He has been invited to give presentations at various scientific meetings and workshops

2021. He is the Editor-in-Chief of the Journal of Advances in Information Technology (JAIT)