scholarly journals Deep COVID 19: Deep Learning for COVID 19 Detection from X ray Images

Author(s):  
Ahmed Hashem El Fiky ◽  

The COVID-19 will take place for the first time in December 2019 in Wuhan, China. After that, the virus spread all over the world, with over 4.7 million confirmed cases and over 315000 deaths as of the time of writing this report. Radiologists can employ machine learning algorithms developed on radiography pictures as a decision support mechanism to help them speed up the diagnostic process. The goal of this study is to conduct a quantitative evaluation of six off-the-shelf convolutional neural networks (CNNs) for COVID-19 X-ray image analysis. Due to the limited amount of images available for analysis, the CNN transfer learning approach was used. We also developed a simple CNN architecture with a modest number of parameters that does a good job of differentiating COVID-19 from regular X-rays. in this paper, we are used large dataset which contained CXR images of normal patients and patients with COVID-19. the number of CXR images for normal patients are 10,192 image and the number of CXR images for COVID-19 patients are 3,616 images. The results of experiments show the effectiveness and robustness of Deep-COVID-19 and pretrained models like VGG16, VGG19, and MobileNets. Our proposed Model Deep-COVID-19 achieved over 94.5% accuracy.

Author(s):  
Taban Majeed ◽  
Rasber Rashid ◽  
Dashti Ali ◽  
Aras Asaad

AbstractThe Covid-19 first occurs in Wuhan, China in December 2019. After that the virus spread all around the world and at the time of writing this paper the total number of confirmed cases are above 4.7 million with over 315000 deaths. Machine learning algorithms built on radiography images can be used as a decision support mechanism to aid radiologists to speed up the diagnostic process. The aim of this work is to conduct a critical analysis to investigate the applicability of convolutional neural networks (CNNs) for the purpose of COVID-19 detection in chest X-ray images and highlight the issues of using CNN directly on the whole image. To achieve this task, we first use 12-off-the-shelf CNN architectures in transfer learning mode on 3 publicly available chest X-ray databases together with proposing a shallow CNN architecture in which we train it from scratch. Chest X-ray images fed into CNN models without any preprocessing to follow the many of researches using chest X-rays in this manner. Next, a qualitative investigation performed to inspect the decisions made by CNNs using a technique known as class activation maps (CAM). Using CAMs, one can map the activations contributed most to the decision of CNNs back to the original image to visualize the most discriminating regions on the input image.We conclude that CNN decisions should not be taken into consideration, despite their high classification accuracy, until clinicians can visually inspect, and approve, the region(s) of the input image used by CNNs that lead to its prediction.


Atmosphere ◽  
2019 ◽  
Vol 10 (4) ◽  
pp. 169 ◽  
Author(s):  
Mahbubur Rahman ◽  
Pasan Hettiarachchi ◽  
Vernon Cooray ◽  
Joseph Dwyer ◽  
Vladimir Rakov ◽  
...  

We present observations of X-rays from laboratory sparks created in the air at atmospheric pressure by applying an impulse voltage with long (250 µs) rise-time. X-ray production in 35 and 46 cm gaps for three different electrode configurations was studied. The results demonstrate, for the first time, the production of X-rays in gaps subjected to switching impulses. The low rate of rise of the voltage in switching impulses does not significantly reduce the production of X-rays. Additionally, the timing of the X-ray occurrence suggests the possibility that the mechanism of X-ray production by sparks is related to the collision of streamers of opposite polarity.


1996 ◽  
Vol 11 (5) ◽  
pp. 1169-1178 ◽  
Author(s):  
Kentaro Suzuya ◽  
Michihiro Furusaka ◽  
Noboru Watanabe ◽  
Makoto Osawa ◽  
Kiyohito Okamura ◽  
...  

Mesoscopic structures of SiC fibers produced from polycarbosilane by different methods were studied by diffraction and small-angle scattering of neutrons and x-rays. Microvoids of a size of 4–10 Å in diameter have been observed for the first time by neutron scattering in a medium momentum transfer range (Q = 0.1–1.0 Å−1). The size and the volume fraction of β–SiC particles were determined for fibers prepared at different heat-treatment temperatures. The results show that wide-angle neutron scattering measurements are especially useful for the study of the mesoscopic structure of multicomponent materials.


Author(s):  
Dipayan Das ◽  
KC Santosh ◽  
Umapada Pal

Abstract Since December 2019, the Coronavirus Disease (COVID-19) pandemic has caused world-wide turmoil in less than a couple of months, and the infection, caused by SARS-CoV-2, is spreading at an unprecedented rate. AI-driven tools are used to identify Coronavirus outbreaks as well as forecast their nature of spread, where imaging techniques are widely used, such as CT scans and chest X-rays (CXRs). In this paper, motivated by the fact that X-ray imaging systems are more prevalent and cheaper than CT scan systems, a deep learning-based Convolutional Neural Network (CNN) model, which we call Truncated Inception Net, is proposed to screen COVID-19 positive CXRs from other non-COVID and/or healthy cases. To validate our proposal, six different types of datasets were employed by taking the following CXRs: COVID-19 positive, Pneumonia positive, Tuberculosis positive, and healthy cases into account. The proposed model achieved an accuracy of 99.96% (AUC of 1.0) in classifying COVID- 19 positive cases from combined Pneumonia and healthy cases. Similarly, it achieved an accuracy of 99.92% (AUC of 0.99) in classifying COVID-19 positive cases from combined Pneumonia, Tuberculosis and healthy CXRs. To the best of our knowledge, as of now, the achieved results outperform the existing AI-driven tools for screening COVID-19 using CXRs.


1987 ◽  
Vol 93 ◽  
pp. 485-485
Author(s):  
H. Steinle ◽  
W. Pietsck

AbstractDuring the August 1983 outburst of the old nova GK Persei observations with EXOSAT showed for the first time a 351 second periodicity in X-rays.Our fast photometry (U(B)V with 25 sec time resolution) was made at the end of the outburst in the nights of September 29 , and October 1–3 , using the 2.2 meter telescope at Calar Alto (Spain).Optical variations up to 10% in U and 4% in V with periodicities in the range 350 to 360 seconds were found, lasting only for few cycles.A comparison with the extrapolated prediction of the X-ray maxima did not show a coincidence, but rather an anticoincidence in several cases. This supports a model of reprocessed X-rays at the inner edge of an accretion disk.


2019 ◽  
Vol 486 (4) ◽  
pp. 4863-4879 ◽  
Author(s):  
Ali Takey ◽  
Florence Durret ◽  
Isabel Márquez ◽  
Amael Ellien ◽  
Mona Molham ◽  
...  

ABSTRACT We present X-ray and optical properties of the optically confirmed galaxy cluster sample from the 3XMM/SDSS Stripe 82 cluster survey. The sample includes 54 galaxy clusters in the redshift range of 0.05–1.2, with a median redshift of 0.36. We first present the X-ray temperature and luminosity measurements that are used to investigate the X-ray luminosity–temperature relation. The slope and intercept of the relation are consistent with those published in the literature. Then, we investigate the optical properties of the cluster galaxies including their morphological analysis and the galaxy luminosity functions (GLFs). The morphological content of cluster galaxies is investigated as a function of cluster mass and distance from the cluster centre. No strong variation of the fraction of early- and late-type galaxies with cluster mass is observed. The fraction of early-type galaxies as a function of cluster radius varies as expected. The individual GLFs of red sequence galaxies were studied in the five ugriz bands for 48 clusters. The GLFs were then stacked in three mass bins and two redshift bins. Twenty clusters of the present sample are studied for the first time in X-rays, and all are studied for the first time in the optical range. Altogether, our sample appears to have X-ray and optical properties typical of ‘average’ cluster properties.


Proceedings ◽  
2019 ◽  
Vol 17 (1) ◽  
pp. 9
Author(s):  
Ka-Wah Wong ◽  
Rodrigo S. Nemmen ◽  
Jimmy A. Irwin ◽  
Dacheng Lin

The nearby M87 hosts an exceptional relativistic jet. It has been regularly monitored in radio to TeV bands, but little has been done in hard X-rays ≳10 keV. For the first time, we have successfully detected hard X-rays up to 40 keV from its X-ray core with joint Chandra and NuSTAR observations, providing important insights to the X-ray origins: from the unresolved jet or the accretion flow. We found that the hard X-ray emission is significantly lower than that predicted by synchrotron self-Compton models introduced to explain very-high-energy γ -ray emission above a GeV. We discuss recent models to understand these high energy emission processes.


2018 ◽  
Vol 619 ◽  
pp. A16
Author(s):  
C. Vignali ◽  
P. Severgnini ◽  
E. Piconcelli ◽  
G. Lanzuisi ◽  
R. Gilli ◽  
...  

Context. The search for heavily obscured active galactic nuclei has been revitalized in the last five years by NuSTAR, which has provided a good census and spectral characterization of a population of such objects, mostly at low redshift, thanks to its enhanced sensitivity above 10 keV compared to previous X-ray facilities, and its hard X-ray imaging capabilities. Aims. We aim at demonstrating how NGC 2785, a local (z = 0.009) star-forming galaxy, is responsible, in virtue of its heavily obscured active nucleus, for significant contamination in the non-imaging BeppoSAX/PDS data of the relatively nearby (≈17′) quasar IRAS 09104+4109 (z = 0.44), which was originally mis-classified as Compton thick. Methods. We analyzed ≈71 ks NuSTAR data of NGC 2785 using the MYTorus model and provided a physical description of the X-ray properties of the source for the first time. Results. We found that NGC 2785 hosts a heavily obscured (NH ≈ 3 × 1024 cm−2) nucleus. The intrinsic X-ray luminosity of the source, once corrected for the measured obscuration (L2−10keV ≈ 1042 erg s−1), is consistent within a factor of a few with predictions based on the source mid-infrared flux using widely adopted correlations from the literature. Conclusions. Based on NuSTAR data and previous indications from the Neil Gehrels Swift Observatory (BAT instrument), we confirm that NGC 2785, because of its hard X-ray emission and spectral shape, was responsible for at least one third of the 20–100 keV emission observed using the PDS instrument onboard BeppoSAX, originally completely associated with IRAS 09104+4109. Such emission led to the erroneous classification of this source as a Compton-thick quasar, while it is now recognized as Compton thin.


2016 ◽  
Vol 23 (6) ◽  
pp. 1333-1347 ◽  
Author(s):  
John P. Sutter ◽  
Simon G. Alcock ◽  
Yogesh Kashyap ◽  
Ioana Nistea ◽  
Hongchang Wang ◽  
...  

Beam shaping is becoming increasingly important for synchrotron X-ray applications. Although routine for visible light lasers, this is challenging for X-rays due to the limited source coherence and extreme optical tolerances required for the shaping mirrors. In deliberate defocusing, even surface errors <5 nm r.m.s. introduce damagingly large striations into the reflected beam. To counteract such problems, surface modifications with alternating concave and convex curvature on equal segments were polished onto the surface of non-active mirrors of fixed curvature. Such optics are useful for providing a fixed size of X-ray beam, but do not provide the adaptability required by many experiments. In contrast, deformable piezo bimorph mirrors permit a continuous range of X-ray beam sizes and shapes. A new theory is developed for applying non-periodic modifications of alternating curvature to optical surfaces. The position and length of the segments may be freely chosen. For the first time, surface modifications of alternating curvature are applied to bimorph mirrors to generate non-Gaussian X-ray beam profiles of specified width. The new theory's freedom is exploited to choose the segments to match the polishing errors of medium wavelength (>10 mm) and the piezos' influence on the mirror's figure. Five- and seven-segment modifications of alternating curvature are calculated and verified by visible light and X-ray metrology. The latter yields beam profiles with less striation than those made by defocusing. Remaining beam striations are explained by applying geometrical optics to the deviations from the ideal surface modifications of alternating curvature.


2021 ◽  
Vol 11 (21) ◽  
pp. 10301
Author(s):  
Muhammad Shoaib Farooq ◽  
Attique Ur Rehman ◽  
Muhammad Idrees ◽  
Muhammad Ahsan Raza ◽  
Jehad Ali ◽  
...  

COVID-19 has been difficult to diagnose and treat at an early stage all over the world. The numbers of patients showing symptoms for COVID-19 have caused medical facilities at hospitals to become unavailable or overcrowded, which is a major challenge. Studies have recently allowed us to determine that COVID-19 can be diagnosed with the aid of chest X-ray images. To combat the COVID-19 outbreak, developing a deep learning (DL) based model for automated COVID-19 diagnosis on chest X-ray is beneficial. In this research, we have proposed a customized convolutional neural network (CNN) model to detect COVID-19 from chest X-ray images. The model is based on nine layers which uses a binary classification method to differentiate between COVID-19 and normal chest X-rays. It provides COVID-19 detection early so the patients can be admitted in a timely fashion. The proposed model was trained and tested on two publicly available datasets. Cross-dataset studies are used to assess the robustness in a real-world context. Six hundred X-ray images were used for training and two hundred X-rays were used for validation of the model. The X-ray images of the dataset were preprocessed to improve the results and visualized for better analysis. The developed algorithm reached 98% precision, recall and f1-score. The cross-dataset studies also demonstrate the resilience of deep learning algorithms in a real-world context with 98.5 percent accuracy. Furthermore, a comparison table was created which shows that our proposed model outperforms other relative models in terms of accuracy. The quick and high-performance of our proposed DL-based customized model identifies COVID-19 patients quickly, which is helpful in controlling the COVID-19 outbreak.


Sign in / Sign up

Export Citation Format

Share Document