XRayCovid19 / Git / Diff of /README.md

Datasets:

GeorgeSullivan/

XRayCovid19

Downloads: 1

Diff of /README.md [addc8c] .. [ed7178]

Switch to unified view


# COVID-19 xray dataset


<img src="https://github.com/v7labs/covid-19-xray-dataset/blob/master/media/covid-chest-xray-cover.jpg?raw=true"/>

[BROWSE & DOWNLOAD THE DATASET ON V7 DARWIN HERE](https://darwin.v7labs.com/v7-labs/covid-19-chest-x-ray-dataset)

or run the following command from [Darwin-py](https://v7labs.github.io/darwin-py/) to download the latest version

...
- A tag for the type of pneumonia (viral, bacterial, fungal, healthy/none)
- If the patient has COVID-19, additional tags stating age, sex, temperature, location, intubation status, ICU admission, and patient outcome.

Lung annotations are polygons following pixel-level boundaries. These can be exported as `COCO`, `VOC`, or `Darwin JSON` formats. Each annotation file contains a URL to the original full resolution image, as well as a reduced size thumbnail.

<img src="https://github.com/v7labs/covid-19-xray-dataset/blob/master/media/example_1.png?raw=true" />

**LUNG SEGMENTATION NOTES**: Lung segmentations in this dataset include most of the heart, revealing lung opacities behind the heart which may be relevant for assessing the severity of viral pneumonia. Uniformly shaped lungs also de-couples the shape and content within the left lung from the size of the heart.

The lower-most part of the lungs is defined by the extent of the diaphragm, where visible. If the back of the lungs is clearly visible through the diaphragm it is also included up until the lower-most visible part of the lungs.

<img src="https://github.com/v7labs/covid-19-xray-dataset/blob/master/media/Jun-17-2020_15-21-23.gif?raw=true"/>

Lung segmentations were performed by human annotators using [Auto-Annotate](https://www.v7labs.com/automated-annotation), adjusted, and reviewed by humans.

Other important notes:

...
- Lateral x-rays do not contain lung segmentations. They have classification tags, but should be ignored if you are working with detection-based networks.
- There are 63 axial CT scan slices left un-labelled with masks (although they contain tags) as a way of maintaining integrity to one of the source datasets. We encourage discarding these when performing x-ray analysis.
- Portable x-ray images are of significant lower quality than others. Be aware that they correlate highly with severe conditions. Classification models will bias portable x-ray images with diseases like COVID-19.
- Medical instruments like pacemakers and markup that overlap the lungs are masked with an "Ignore" class. We encourage masking these out when performing lung analysis as they correlated strongly with sick patients. Intubation instruments are not removed if smaller/thinner than 1cm.

<img src="https://github.com/v7labs/covid-19-xray-dataset/blob/master/media/example_2.png?raw=true" />
<img src="https://github.com/v7labs/covid-19-xray-dataset/blob/master/media/example_3.png?raw=true" />
<img src="https://github.com/v7labs/covid-19-xray-dataset/blob/master/media/example_4.png?raw=true" />



You may also use the `Ignore` class to filter out images with occluding markups or large medical instruments.

You can browse the available images and filter them by tag or annotation class using the right-sidebar as seen below. Follow this [Link](https://darwin.v7labs.com/v7-labs/covid-19-chest-x-ray-dataset) to access the interactive dataset.

...

Prof. Paolo Spagnolo

	a/README.md		b/README.md
1	# COVID-19 xray dataset	1	# COVID-19 xray dataset
2		2
3		3
4	<img src="https://github.com/v7labs/covid-19-xray-dataset/blob/master/media/covid-chest-xray-cover.jpg"/>	4	<img src="https://github.com/v7labs/covid-19-xray-dataset/blob/master/media/covid-chest-xray-cover.jpg?raw=true"/>
5		5
6	[BROWSE & DOWNLOAD THE DATASET ON V7 DARWIN HERE](https://darwin.v7labs.com/v7-labs/covid-19-chest-x-ray-dataset)	6	[BROWSE & DOWNLOAD THE DATASET ON V7 DARWIN HERE](https://darwin.v7labs.com/v7-labs/covid-19-chest-x-ray-dataset)
7		7
8	or run the following command from [Darwin-py](https://v7labs.github.io/darwin-py/) to download the latest version	8	or run the following command from [Darwin-py](https://v7labs.github.io/darwin-py/) to download the latest version
9		9
	...		...
27	- A tag for the type of pneumonia (viral, bacterial, fungal, healthy/none)	27	- A tag for the type of pneumonia (viral, bacterial, fungal, healthy/none)
28	- If the patient has COVID-19, additional tags stating age, sex, temperature, location, intubation status, ICU admission, and patient outcome.	28	- If the patient has COVID-19, additional tags stating age, sex, temperature, location, intubation status, ICU admission, and patient outcome.
29		29
30	Lung annotations are polygons following pixel-level boundaries. These can be exported as `COCO`, `VOC`, or `Darwin JSON` formats. Each annotation file contains a URL to the original full resolution image, as well as a reduced size thumbnail.	30	Lung annotations are polygons following pixel-level boundaries. These can be exported as `COCO`, `VOC`, or `Darwin JSON` formats. Each annotation file contains a URL to the original full resolution image, as well as a reduced size thumbnail.
31		31
32	<img src="https://github.com/v7labs/covid-19-xray-dataset/blob/master/media/example_1.png" />	32	<img src="https://github.com/v7labs/covid-19-xray-dataset/blob/master/media/example_1.png?raw=true" />
33		33
34	LUNG SEGMENTATION NOTES: Lung segmentations in this dataset include most of the heart, revealing lung opacities behind the heart which may be relevant for assessing the severity of viral pneumonia. Uniformly shaped lungs also de-couples the shape and content within the left lung from the size of the heart.	34	LUNG SEGMENTATION NOTES: Lung segmentations in this dataset include most of the heart, revealing lung opacities behind the heart which may be relevant for assessing the severity of viral pneumonia. Uniformly shaped lungs also de-couples the shape and content within the left lung from the size of the heart.
35		35
36	The lower-most part of the lungs is defined by the extent of the diaphragm, where visible. If the back of the lungs is clearly visible through the diaphragm it is also included up until the lower-most visible part of the lungs.	36	The lower-most part of the lungs is defined by the extent of the diaphragm, where visible. If the back of the lungs is clearly visible through the diaphragm it is also included up until the lower-most visible part of the lungs.
37		37
38	![media/Jun-17-2020_15-21-23.gif](media/Jun-17-2020_15-21-23.gif)	38	<img src="https://github.com/v7labs/covid-19-xray-dataset/blob/master/media/Jun-17-2020_15-21-23.gif?raw=true"/>
39		39
40	Lung segmentations were performed by human annotators using [Auto-Annotate](https://www.v7labs.com/automated-annotation), adjusted, and reviewed by humans.	40	Lung segmentations were performed by human annotators using [Auto-Annotate](https://www.v7labs.com/automated-annotation), adjusted, and reviewed by humans.
41		41
42	Other important notes:	42	Other important notes:
43		43
	...		...
45	- Lateral x-rays do not contain lung segmentations. They have classification tags, but should be ignored if you are working with detection-based networks.	45	- Lateral x-rays do not contain lung segmentations. They have classification tags, but should be ignored if you are working with detection-based networks.
46	- There are 63 axial CT scan slices left un-labelled with masks (although they contain tags) as a way of maintaining integrity to one of the source datasets. We encourage discarding these when performing x-ray analysis.	46	- There are 63 axial CT scan slices left un-labelled with masks (although they contain tags) as a way of maintaining integrity to one of the source datasets. We encourage discarding these when performing x-ray analysis.
47	- Portable x-ray images are of significant lower quality than others. Be aware that they correlate highly with severe conditions. Classification models will bias portable x-ray images with diseases like COVID-19.	47	- Portable x-ray images are of significant lower quality than others. Be aware that they correlate highly with severe conditions. Classification models will bias portable x-ray images with diseases like COVID-19.
48	- Medical instruments like pacemakers and markup that overlap the lungs are masked with an "Ignore" class. We encourage masking these out when performing lung analysis as they correlated strongly with sick patients. Intubation instruments are not removed if smaller/thinner than 1cm.	48	- Medical instruments like pacemakers and markup that overlap the lungs are masked with an "Ignore" class. We encourage masking these out when performing lung analysis as they correlated strongly with sick patients. Intubation instruments are not removed if smaller/thinner than 1cm.
49		49
50	![media/example_2.png](media/example_2.png)	50	<img src="https://github.com/v7labs/covid-19-xray-dataset/blob/master/media/example_2.png?raw=true" />
51		51	<img src="https://github.com/v7labs/covid-19-xray-dataset/blob/master/media/example_3.png?raw=true" />
52	![media/example_3.png](media/example_3.png)	52	<img src="https://github.com/v7labs/covid-19-xray-dataset/blob/master/media/example_4.png?raw=true" />
53		53
54	![media/example_4.png](media/example_4.png)
55		54
56	You may also use the `Ignore` class to filter out images with occluding markups or large medical instruments.	55	You may also use the `Ignore` class to filter out images with occluding markups or large medical instruments.
57		56
58	You can browse the available images and filter them by tag or annotation class using the right-sidebar as seen below. Follow this [Link](https://darwin.v7labs.com/v7-labs/covid-19-chest-x-ray-dataset) to access the interactive dataset.	57	You can browse the available images and filter them by tag or annotation class using the right-sidebar as seen below. Follow this [Link](https://darwin.v7labs.com/v7-labs/covid-19-chest-x-ray-dataset) to access the interactive dataset.
59		58
	...		...
105		104
106	Prof. Paolo Spagnolo	105	Prof. Paolo Spagnolo