Content uploaded by Priya Ranjan
Author content
All content in this area was uploaded by Priya Ranjan on Apr 12, 2020
Content may be subject to copyright.
A preview of the PDF is not available
Pneumonia Identification in Chest X-Ray Images Using EMD
Figures
Figures - uploaded by Priya Ranjan
Author content
All figure content in this area was uploaded by Priya Ranjan
Content may be subject to copyright.
... The X-ray technique has been used to diagnose abnormalities in the regions of the human body like the chest, skull, bones, teeth, etc. For many decades, medical practitioners have used the X-ray method to analyze and explore the various abnormalities in human body organs [6]. Many studies revealed that X-rays are a cost-effective method for disease diagnosis while revealing pathological alterations with their economic efficiency and non-invasive properties [7]. ...
... Rights reserved. Figures 5,6,7,8 and 9 show the outcome of accuracy, precision, recall, specificity, and F1-score parameters, respectively. The results show that performances were measured by applying five classifiers (machine and deep learning), ANN, SVM, RF, CNN, and LSTM on the lung dataset. ...
Heart disease has been the leading cause of mortality worldwide in the recent decade. Since 2019, new lung-related infections have increased heart attack mortality. To minimize mortality, a unified framework is needed to predict early diagnosis utilizing different patient data. Existing methods failed to produce automatic solutions on the unified approach to cardiac problems considering lung infections. A unique automated disease diagnosis and classification framework using patient chest X-ray images and Electrocardiogram (ECG) signal is proposed. This integrated framework is unique in diagnosing and monitoring lung disease and cardiac problems utilizing patient X-ray and ECG. The proposed system is called the Deep Intelligence Framework for Diseases Detection (DIFDD). DIFDD procedures include pre-processing, automated feature extraction, and classification. An effective pre-processing method is designed to improve X-ray and ECG data. The 2D and 1D Convolutional Neural Network (CNN) techniques are proposed to extract automated features from pre-processed X-ray and ECG data. According to feature learning, automated detection uses several classifiers. Based on classifier results, a consolidated approach is presented for medical judgment on the patient's health for suitable treatment. The simulation results using the synthetic dataset revealed the efficiency of the proposed DIFDD model over the existing methods. The overall accuracy of the DIFDD model is improved by 1.5% and computational overhead is reduced by 13.56%.
... This experiment is conducted to compare the performance of the proposed ViBaNet with other methods to show the dominance of our method. Consequently, we compared the proposed ViBaNet with other techniques [40,41] that are utilized for the identification of viral and bacterial pneumonia patients. The data of both diseases are insufficient; therefore, we employed the data augmentation technique to enlarge the amount of data and avoid the common problem i.e., the overfitting problem. ...
... As depicted in Figure 3., we have evaluated the efficacy of other methods with our ViBaNet method. Figure 3 shows that Xianghong et al. [40] obtained an accuracy of 80.4%, precision of 88.8%, and recall of 77.5% while A. Khatri et al. [41] used the EMD approach and achieved an accuracy of 83.3%, precision of 80.0%, and recall of 89.5%. The proposed ViBaNet obtained an accuracy, precision, and recall of 92.56%, 95.65%, and 96.35%, respectively as illustrated in Fig 3. ...
Viral and bacterial pneumonia mostly occurs in lungs of the humans and are life-threatening diseases if timely treatment is ignored. In this regard, age is a very crucial aspect as the effects are different in the age of different people. Most commonly, infants, as well as aged people are at risk due to these diseases as their lungs are badly affected. It is a really challenging, and time-consuming task for diagnostician to inspect the lung’s radiographic scans and diagnose pneumonia. Binary classification such as normal vs pneumonia or normal vs bacterial pneumonia is simple, however, detecting bacterial and viral pneumonia is a tricky and difficult task. Therefore, we implemented a deep-learning method to identify pneumonia victims effectively by employing lung X-rays to avoid wrong decisions by radiologists. In the present study, we proposed a novel deep-learning technique, ViBaNet, which is based on the customized residual neural network to investigate the validity of ViBaNet to diagnose complicated pneumonia disorders. We conducted experiments employing the uncustomed and ViBaNet and evaluated the efficacy of these techniques. The novel ViBaNet obtained an accuracy, precision, recall, and F1-score of 92.56%, 95.65%, 96.35%, and 96%, respectively. The above-mentioned analysis results provide evidence that the ViBaNet is effective to be utilized to identify pneumonia individuals. Moreover, the comparison of ViBaNet with other techniques has shown superior performance results using the augmented data.
... The fo llowing techniques utilized the distance of two probability distributions over the region of interest. However, despite concentrating on specific regions of the image, there is enough need to increase the accuracy in specific applications of greater significance [7]. Contrary to the implementation of lightweight procedures as discussed above, studies were done to assess the performance of neural networks based on classification accuracy [8]. ...
... In [49], authors introduced a computed tomography(CT)-based network model known as multi-scale heterogeneous 3D CNN (MSH-CNN) for detection and prediction. In [22], authors have introduced the concept of the distance earth mover's to recognize the infected lungs from the normal ones for pneumonia. Authors of [8] used majority voting to combine the results of various network models to achieve the better results for prediction and detection. ...
Pneumonia is an infection that inflames the air sacs in lungs and is one of the prime causes of deaths under the age of five, all over the world. Moreover, sometimes it became quite difficult to detect and diagnose pneumonia by just looking at the chest plain X-ray images. Therefore, we propose a hybrid deep convolutional neural network model (HDCNN) for improved diagnosis of pneumonia, to simplify the detection for medical practitioners and specialists. In this proposed model, image preprocessing is performed using Student's t distribution, a compact probability density function (cPDF), for better sampling and segregation between the healthy and infected part of lungs, to improve the predictions. Further, a hybrid deep convolutional neural network model is built to extract image features by fine-tuning the pretrained models, viz. Resnet-50, EfficientNet, VGG-16, MobileNetV2 and DenseNet to achieve better results of diagnosis. The proposed hybrid model is analyzed using Grad-CAM visualization, which produces a course localization map, highlighting the infected region in the image used for prediction. The proposed hybrid model is evaluated based on governing parameters, viz. precision, recall, F1-score and accuracy. The results show our proposed model achieves precision of 97.47%, recall 98.09%, F1-score of 97.77% and overall accuracy of 97.69% as compared to other existing models.
... After the image has been preprocessed, it is scaled and normalised by intensity to provide a set of uniform lung sizes and shapes. The existing and forthcoming issues in medical image processing were identified [21]. A number of studies regarding the usage of deep learning methods in the finding of various diseases are highlighted in [22]. ...
Artificial intelligence and machine learning will be the driving forces behind the next computing revolution. These technologies rely on the ability to identify trends from historical information and predict future outcomes. One of the best machine learning techniques, deep learning is employed in a variety of applications, including object recognition, picture categorization, image analysis, and clinical archives. Image and video data are necessary for both diagnosing the patient's illness and determining its severity. Convolutional neural networks are efficient gears for digital picture classification and image understanding. The production of medical photographs has ex-ponentially increased as a result of the proliferation of digital devices and the development of camera technology, which creates Bigdata. Massive, difficult-to-manage volumes of structured, unstructured data are referred to as "Big data". The more data processed for analysis, the greater will be the analytical accuracy and also the greater would be the confidence in our decisions based on the analytical findings. In this paper, we proposed a novel method for early detection of pneumonia disease using deep learning techniques along with the big data storage and big data analytics to achieve more better performance. The results show that, the model achieved 91.16% of accuracy and 93.22% of F1-score.
... Among the studies listed in the [55] achieved an accuracy of 84.5% using a dataset similar to the previous studies. Khatri, Archit [56] achieved a recall of 89.5%, precision of 80%, accuracy of 83.3%, and an f1-score of 84.75% on a dataset with bacterial and viral pneumonia cases. Xianghong Gu et al. [57] obtained an accuracy of 80.4% on a similar dataset. ...
Pneumonia is a significant global health concern that can lead to severe and sometimes fatal consequences. Timely identification and classification of pneumonia can substantially improve patient outcomes. However, the disclosure of sensitive medical data for diagnostic purposes raises important issues regarding patient privacy and data security. The Federated Multi-Party Computing (FMPC) technique has emerged as a promising solution to these challenges, enabling multiple parties to collaborate and compute a function over their private data while maintaining data privacy. This paper presents a novel Internet of Things (IoT)-enabled FMPC framework that significantly enhances the accuracy of pneumonia categorization while preserving patient privacy and data security. Using publicly available Kaggle datasets, the efficacy of the proposed strategy is assessed, and a comparison with contemporary deep learning models is made. The study demonstrates the remarkable efficacy of the IoT-enabled FMPC approach in pneumonia recognition within the industry 5.0 landscape. With 96.67% accuracy achieved for the unbalanced dataset and 97.84% accuracy for the balanced dataset, the results of the proposed algorithm demonstrate the potential for improvement. This approach ensures the privacy of sensitive medical information, aligning with the core principles of Industry 5.0 that emphasize the harmonious integration of advanced technologies and human-centric values.
... It happens when a person has Pneumonia. There are two different kinds of Pneumonia, bacterial and viral, yet the X-ray patterns of each look highly similar [15]. That requires guidance from computer-aided diagnostic tools. ...
This review paper examines the current state of lung disease diagnosis based on deep learning (DL) methods. Lung diseases, such as Pneumonia, TB, Covid-19, and lung cancer, are significant causes of morbidity and mortality worldwide. Accurate and timely diagnosis of these diseases is essential for effective treatment and improved patient outcomes. DL methods, which utilize artificial neural networks to extract features from medical images automatically, have shown great promise in improving the accuracy and efficiency of lung disease diagnosis. This review discusses the various DL methods that have been developed for lung disease diagnosis, including convolutional neural networks (CNNs), deep neural networks (DNNs), and generative adversarial networks (GANs). The advantages and limitations of each method are discussed, along with the types of medical imaging techniques used, such as X-ray and computed tomography (CT). In addition, the review discusses the most commonly used performance metrics for evaluating the performance of DL for lung disease diagnosis: the area under the curve (AUC), sensitivity, specificity, F1-score, accuracy, precision, and the receiver operator characteristic curve (ROC). Moreover, the challenges and limitations of using DL for lung disease diagnosis, including the limited availability of annotated data, the variability in imaging techniques and disease presentation, and the interpretability and generalizability of DL models, are highlighted in this paper. Furthermore, strategies to overcome these challenges, such as transfer learning, data augmentation, and explainable AI, are also discussed. The review concludes with a call for further research to address the remaining challenges and realize DL's full potential for improving lung disease diagnosis and treatment.
... Thus, there is a need to explore how artificial intelligence (AI) can assist in other radiological tasks, such as pneumonia detection. detecting pneumonia on CXRs, with recent models attaining a classification accuracy ranging from 67% to 96% and area under the curve of the receiver operating characteristics graphs (AUROC) ranging from 0.65 to 0.99 [7][8][9][10][11]. Even though CNNs give promising accuracy and AUROC, these models can be further improved through hyperparameter tuning. ...
Background: Pneumonia is an infectious disease that is especially harmful to those with weak immune systems, such as children under the age of 5. While radiologists’ diagnosis of pediatric pneumonia on chest radiographs (CXRs) is often accurate, subtle findings can be missed due to the subjective nature of the diagnosis process. Artificial intelligence (AI) techniques, such as convolutional neural networks (CNNs), can help make the process more objective and precise. However, off-the-shelf CNNs may perform poorly if they are not tuned to their appropriate hyperparameters. Our study aimed to identify the CNNs and their hyperparameter combinations (dropout, batch size, and optimizer) that optimize model performance.
Methodology: Sixty models based on five CNNs (VGG 16, VGG 19, DenseNet 121, DenseNet 169, and InceptionResNet V2) and 12 hyperparameter combinations were tested. Adam, Root Mean Squared Propagation (RmsProp), and Mini-Batch Stochastic Gradient Descent (SGD) optimizers were used. Two batch sizes, 32 and 64, were utilized. A dropout rate of either 0.5 or 0.7 was used in all dropout layers. We used a deidentified CXR dataset of 4200 pneumonia (Figure 1a) and 1600 normal images (Figure 1b). Seventy percent of the CXRs in the dataset were used for training the model, 20% were used for validating the model, and 10% were used for testing the model. All CNNs were trained first on the ImageNet dataset. They were then trained, with frozen weights, on the CXR-containing dataset.
Results: Among the 60 models, VGG-19 (dropout of 0.5, batch size of 32, and Adam optimizer) was the most accurate. This model achieved an accuracy of 87.9%. A dropout of 0.5 consistently gave higher accuracy, area under the receiver operating characteristics curve (AUROC), and area under the precision-recall curve (AUPRC) compared to a dropout of 0.7. The CNNs InceptionResNet V2, DenseNet 169, VGG 16, and VGG 19 significantly outperformed the DenseNet121 CNN in accuracy and AUROC. The Adam and RmsProp optimizer had improved AUROC and AUPRC compared to the SGD optimizer. The batch size had no statistically significant effect on model performance.
Conclusion: We recommend using low dropout rates (0.5) and RmsProp or Adam optimizer for pneumonia-detecting CNNs. Additionally, we discourage using the DenseNet121 CNN when other CNNs are available. Finally, the batch size may be set to any value, dependent on computational resources.
... Toğacar et al. [13] compared the performance of traditional machine learning models for detecting pneumonia with the Redundancy Maximum Relevance (mRMR) minimum feature selection mechanism. Khatri et al. [14] compared CXR pneumonia images using the Earth Mover's Distance (EMD). Teixeira et al. [15] evaluated and described COVID-19 using lung segmentation. ...
Pneumonia is a highly lethal disease, and research on its treatment and early screening tools has received extensive attention from researchers. Due to the maturity and cost reduction of chest X-ray technology, and with the development of artificial intelligence technology, pneumonia identification based on deep learning and chest X-ray has attracted attention from all over the world. Although the feature extraction capability of deep learning is strong, existing deep learning object detection frameworks are based on pre-defined anchors, which require a lot of tuning and experience to guarantee their excellent results in the face of new applications or data. To avoid the influence of anchor settings in pneumonia detection, this paper proposes an anchor-free object detection framework and RSNA dataset based on pneumonia detection. First, a data enhancement scheme is used to preprocess the chest X-ray images; second, an anchor-free object detection framework is used for pneumonia detection, which contains a feature pyramid, two-branch detection head, and focal loss. The average precision of 51.5 obtained by Intersection over Union (IoU) calculation shows that the pneumonia detection results obtained in this paper can surpass the existing classical object detection framework, providing an idea for future research and exploration.
The enormous pressure towards efficient city management has triggered various Smart City initiatives by both government and private sector businesses to invest in Information and Communication Technologies to find sustainable solutions to the diverse opportunities and challenges (e.g., waste management). Several researchers have attempted to define and characterize smart cities and then identify opportunities and challenges in building smart cities. This short article also articulates the ongoing movement of Internet of Things and its relationship to smart cities.
This chapter revises the most important aspects in how computing infrastructures should be configured and intelligently managed to fulfill the most notably security aspects required by Big Data applications. One of them is privacy. It is a pertinent aspect to be addressed because users share more and more personal data and content through their devices and computers to social networks and public clouds. So, a secure framework to social networks is a very hot topic research. This last topic is addressed in one of the two sections of the current chapter with case studies. In addition, the traditional mechanisms to support security such as firewalls and demilitarized zones are not suitable to be applied in computing systems to support Big Data. SDN is an emergent management solution that could become a convenient mechanism to implement security in Big Data systems, as we show through a second case study at the end of the chapter. This also discusses current relevant work and identifies open issues.
To detect pulmonary abnormalities such as Tuberculosis (TB), an automatic analysis and classification of chest radiographs can be used as a reliable alternative to more sophisticated and technologically demanding methods (e.g. culture or sputum smear analysis). In target areas like Kenya TB is highly prevalent and often co-occurring with HIV combined with low resources and limited medical assistance. In these regions an automatic screening system can provide a cost-effective solution for a large rural population. Our completely automatic TB screening system is processing the incoming CXRs (chest X-ray) by applying image preprocessing techniques to enhance the image quality followed by an adaptive segmentation based on model selection. The delineated lung regions are described by a multitude of image features. These characteristics are than optimized by a feature selection strategy to provide the best description for the classifier, which will later decide if the analyzed image is normal or abnormal. Our goal is to find the optimal feature set from a larger pool of generic image features, -used originally for problems such as object detection, image retrieval, etc. For performance evaluation measures such as under the curve (AUC) and accuracy (ACC) were considered. Using a neural network classifier on two publicly available data collections, -namely the Montgomery and the Shenzhen dataset, we achieved the maximum area under the curve and accuracy of 0.99 and 97.03%, respectively. Further, we compared our results with existing state-of-the-art systems and to radiologists' decision.
Wireless Sensor Networks (WSNs) are made up of small low-power nodes that are used in different areas like environment monitoring and several military and civilian applications. But due to its small size and limited energy source, energy efficiency is its main area of concern, and many methods have been developed to improve its network lifetime. Game Theory is being used in WSNs to improve the energy efficiency of a network and its lifetime. Game Theory is suitable for such problems as it can be used in node or network level to encourage the decision-making capabilities of WSNs. This survey paper focuses on different types of clustering protocols designed in WSNs using Game Theory to combat the problem of energy efficiency. In particular, we address the approaches by which Game Theory has been used in WSNs to improve its network lifetime including the games used in each protocol.
The major airway diseases in children (cystic fibrosis, primary cilia dyskinesia, asthma, bronchiolitis obliterans, and Swyer–James) are being discussed in the Chapter Airways.
It is well recognized that a large number of pulmonary diseases are induced by the effects of inhaled particulates. Anthracosis is defined as an asymptomatic, mild form of pneumoconiosis caused by the accumulation of “black carbon” in the lungs due to repeated exposure to air pollution or inhalation of smoke or coal dust particles. Since the human population is progressively exposed to an increasing number and doses of anthropogenic micro and nano particles/compounds, there is a pressing urgency to explore toxicological impact arising from these exposures and the molecular mechanisms driving the body defense or possible related diseases. The toxicity mechanisms are clearly related to chemical composition and physical and surface properties of materials. A combination of synchrotron radiation-based (SR-based) nano X-ray fluorescence (XRF) imaging and soft X-ray microscopy was used to chemically characterize environmental particulates (anthracosis) in lung tissues from urban subjects with the aim of better understanding the complex nature of related lungs' deposits. High-resolution XRF analyses performed at ESRF and Elettra synchrotrons allowed discriminating single particles in the heterogeneous aggregates found in the lung tissue. The small particles have variable composition resulting from the different combinations of Ti with O, K and Si, Al and Si, or Zn and Fe with O. Interestingly, simultaneous absorption and phase contrast images showed the particulate morphology and allowed to predict the presence of very dense nanoparticles or high concentration of heavy elements.
The recent progress of computing, machine learning, and especially deep learning, for image recognition brings a meaningful effect for automatic detection of various diseases from chest X-ray images (CXRs). Here efficiency of lung segmentation and bone shadow exclusion techniques is demonstrated for analysis of 2D CXRs by deep learning approach to help radiologists identify suspicious lesions and nodules in lung cancer patients. Training and validation was performed on the original JSRT dataset (dataset #01), BSE-JSRT dataset, i.e. the same JSRT dataset, but without clavicle and rib shadows (dataset #02), original JSRT dataset after segmentation (dataset #03), and BSE-JSRT dataset after segmentation (dataset #04). The results demonstrate the high efficiency and usefulness of the considered pre-processing techniques in the simplified configuration even. The pre-processed dataset without bones (dataset #02) demonstrates the much better accuracy and loss results in comparison to the other pre-processed datasets after lung segmentation (datasets #02 and #03).
The purpose of our study was to evaluate any differences between lung ultrasonography and chest radiography (CR) images in children with a diagnosis of community-acquired pneumonia (CAP) and, if there are any, to analyze the reasons and possible clinical implications. We reviewed the medical records of patients admitted to the pediatric ward from January 2014 to December 2016 and selected only cases discharged with a diagnosis of CAP who had undergone performed lung ultrasound (LUS) and CR within 24 h of each other. All radiologic and sonographic images of the selected cases were examined blindly by a senior radiologist and a skilled sonographer, respectively, with respect to number, position and size of lung injuries. Of the 47 cases of pneumonia, 28 lung lesions spotted by LUS were undetected by CR. Compared with CR, LUS detects more cases of pneumonia, a greater number of cases of double pneumonia and minimal pleural effusions. LUS should be considered the first-line imaging tool for CAP.