Embedded Files
X-ray and CT datasets for most of the known diseases are accessible from several public and open source organizations. As COVID-19 being a very recent emergence around the world, the availability of labelled COVID-19 positive data is limited.so. datasets used in this paper is a combination of many datasets from different online open-source database platform like Kaggle. Our datasets consist of a total of 20,000 images. These include chests X-ray, CT and other chest pneumonia diseases. The table below gives a detailed description of our datasets.
The covid-19 and normal Xray images are acquired from El-Shafai, Walid, Abd El-Samie and Fathi from the Mendeley webpage, and Bachir from Kaggle. we obtained a huge amount of our normal and covid-19 Xray images from these sources, but the images vary in quality and have many clinical markings and medical instrument projections on x-ray images, Our analysis in this article is based on a download of this dataset made on 26 October 2020.
The CT covid-19 and non-covid-19 images were collected from Kaggle published by Abu Zahid Bin Aziz which was our main source for 4000 covid-19 CT images, all of these CT images are of good quality with no noise in images, but has clinical markings. lastly, some of the CT covid-19 images were also taken from webpage publication by El-Shafai, Walid, Our analysis in this article is based on a download of this dataset made on 15 December 2020.
For our studies on multi labelled data, the CT and X-ray pneumonia images of different viral and bacterial strains are taken from open source publications, firstly the pneumonia X-ray images are taken from Kaggle published by Paul Mooney but once again the images have many clinical markings and have variations in size and quality. CT pneumonia images are collected from Kaggle by Mohamed Hany, these ct images are fairly good in quality but since most of the CT images are a full scan consisting of many discrete slices taken in a helical pattern along the thoracic cavity, many images had to be filtered.
The covid-19 and normal Xray images are acquired from El-Shafai, Walid, Abd El-Samie and Fathi from the Mendeley webpage, and Bachir from Kaggle. we obtained a huge amount of our normal and covid-19 Xray images from these sources, but the images vary in quality and have many clinical markings and medical instrument projections on x-ray images, Our analysis in this article is based on a download of this dataset made on 26 October 2020.
The CT covid-19 and non-covid-19 images were collected from Kaggle published by Abu Zahid Bin Aziz which was our main source for 4000 covid-19 CT images, all of these CT images are of good quality with no noise in images, but has clinical markings. lastly, some of the CT covid-19 images were also taken from webpage publication by El-Shafai, Walid, Our analysis in this article is based on a download of this dataset made on 15 December 2020.
For our studies on multi labelled data, the CT and X-ray pneumonia images of different viral and bacterial strains are taken from open source publications, firstly the pneumonia X-ray images are taken from Kaggle published by Paul Mooney but once again the images have many clinical markings and have variations in size and quality. CT pneumonia images are collected from Kaggle by Mohamed Hany, these ct images are fairly good in quality but since most of the CT images are a full scan consisting of many discrete slices taken in a helical pattern along the thoracic cavity, many images had to be filtered.
DATASET :
The dataset consists of five folders i.e, Binary_X_ray and Binary_CT folders with covid-19 and non-covid-19 images, Multi_ct and Multi_X_ray folders for covid-19, normal and phenomena ct and x-ray images and a Binary_sample folder consisting of 1500 training and 500 test image samples from original data for covid-19 and non-covid-19 conditions respectively for CT and X-ray modes to test the data for deep learning architectures.
Contributors :
Rehnuma Shabnam Ananya
M.Sc. Digital Engineering
Otto von guericke universität
magdeburg,Germany
rehnuma.ananya@st.ovgu.de
Shilpa Babu
M.Sc. Digital Engineering
Otto von guericke universität
magdeburg,Germany
shilpa.babu@st.ovgu.de
Aravind Lade
M.Sc. Digital Engineering
Otto von guericke universität
magdeburg,Germany
aravind.lade@st.ovgu.de
Supervisor:
Prof. Dr.-Ing. Andreas Nürnberger
Data and Knowledge Engineering
Otto von guericke universität
magdeburg,Germany
andreas.nuernberger@ovgu.de
Jiahua Xu
Data and Knowledge Engineering
Otto von guericke universität
magdeburg,Germany
jiahua.xu@med.ovgu.de
M.Sc. Philipp Ernst
Data and Knowledge Engineering
Otto von guericke universität
magdeburg,Germany
philipp.ernst@ovgu.de
Reference :
https://data.mendeley.com/datasets/8h65ywd2jr/3 [ Walid El-Shafai,Fathi Abd El-Samie]
https://www.kaggle.com/khushwantparihar/covid-detectionxray-images [Khushwant Parihar]
https://www.kaggle.com/tawsifurrahman/covid19-radiography-database [Tawsifur Rahman, Dr. Muhammad Chowdhury, Amith Khandakar]
W. Ning, S. Lei, J. Yang et al., “Open resource of clinical data from patients with pneumonia for the prediction of covid-19 outcomes via deep learning,” Nature Biomedical Engineering, 2020 [Wanshan Ning, Shijun Lei, Jingjing Yang, Yukun Cao]
https://www.kaggle.com/paultimothymooney/chest-xray-pneumonia [Paul Mooney ]
“Chestx-ray8: Hospital-scale chest xray database and benchmarks on weakly-supervised classification and localization of common thorax diseases " [Xiaosong Wang, Yifan Peng, Le Lu, Zhiyong Lu ]
Identifying medical diagnoses and treatable diseases by image-based deep learning [Daniel S. Kermany, Michael Goldbaum,Wenjia Cai]
: https: //www.kaggle.com/mohamedhanyyy/chest-ctscan-images [Mohamed Hany ]
Page updated
Google Sites
Report abuse