Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom

Improvements to dataset tables #649

Merged
merged 2 commits into from
Jul 7, 2022
Merged

Improvements to dataset tables #649

merged 2 commits into from
Jul 7, 2022

Conversation

adamjstewart
Copy link
Collaborator

Various improvements to dataset tables in the docs. Will comment in-line.

@github-actions github-actions bot added the documentation Improvements or additions to documentation label Jul 2, 2022
adamjstewart
adamjstewart commented Jul 2, 2022
View reviewed changes
docs/api/datasets.rst
@@ -164,8 +164,8 @@ Kenya Crop Type

.. autoclass:: CV4AKenyaCropType

Deep Globe Land Cover Challenge
Copy link
Collaborator Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Name was too long, cut off in miniviewer.

docs/api/geo_datasets.csv
`Aster Global DEM`_,Mask,Aster,"3,601x3,601",30
`Canadian Building Footprints`_,Labels,Generated,,
Copy link
Collaborator Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

"Labels" is ambiguous. When I hear "labels" I think classification or regression. I think "geometries" is more clear for this dataset. Also, "generated" doesn't tell me anything.

Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

"generated" to me is model output vs. manual interpretation

docs/api/geo_datasets.csv
`Aster Global DEM`_,Mask,Aster,"3,601x3,601",30
`Canadian Building Footprints`_,Labels,Generated,,
`Chesapeake Land Cover`_,"Imagery, Labels",,,1
Copy link
Collaborator Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I think "mask" is better than "labels" for semantic segmentation problems

Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Would go with "Masks"

docs/api/geo_datasets.csv
`Canadian Building Footprints`_,Labels,Generated,,
`Chesapeake Land Cover`_,"Imagery, Labels",,,1
`Global Mangrove Distribution`_,Mask,Generated,,3
`Cropland Data Layer`_,Labels,Aerial,,
Copy link
Collaborator Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Imagery is created using Landsat, so same resolution as Landsat

docs/api/geo_datasets.csv
`EU-DEM`_,Mask,"Aster, SRTM, Russian Topomaps",-,25
`GBIF`_,Points,Citizen Scientists,-,-
`GlobBiomass`_,Mask,Landsat,"45,000x45,000",100
`iNaturalist`_,Points,Citizen Scientists,-,-
Copy link
Collaborator Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

iNaturalist was missing from the table.

For GBIF/EDDMapS/iNaturalist I used "points" for the dataset type and "citizen scientists" for the label sources

docs/api/non_geo_datasets.csv
@@ -1,34 +1,34 @@
Dataset,Task,Source,# Samples,# Classes,Size (px),Resolution (m),Bands
`ADVANCE`_,C,"Google Earth, Freesound","5,075",13,512x512,0.5,RGB
`Benin Cashew Plantations`_,S,Airbus Pléiades,70,6,"1,186x1,122",0.5,MSI
`Benin Cashew Plantations`_,S,Airbus Pléiades,70,6,"1,122x1,186",10,MSI
Copy link
Collaborator Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

If you run gdalinfo on the files in @calebrob6's VM, the pixel size is 10, not 0.5. Also, dimensions are reversed.

docs/api/non_geo_datasets.csv
`ETCI2021 Flood Detection`_,S,Sentinel-1,"66,810",2,256x256,5--20,SAR
`EuroSAT`_,C,Sentinel-2,"27,000",10,64x64,10,MSI
`FAIR1M`_,OD,Gaofen/Google Earth,"15,000",37,"1,024x1,024",0.3--0.8,RGB
`Forest Damage`_,OD,Drone imagery,"1,543",4,"1,500x1,500",,RGB
`GID-15`_,S,Gaofen-2,150,15,"6,800x7,200",3,RGB
`IDTReeS`_,"OD,C",Aerial,591,33,200x200,0.1--1,RGB
`Inria Aerial Image Labeling`_,S,Aerial,360,,"5,000x5,000",0.3,RGB
`Inria Aerial Image Labeling`_,S,Aerial,360,2,"5,000x5,000",0.3,RGB
Copy link
Collaborator Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Binary semantic segmentation so 2 classes

docs/api/non_geo_datasets.csv
`Deep Globe Land Cover Challenge`_,S,DigitalGlobe +Vivid,803,7,"2,448x2,448",0.5,RGB
`DFC2022`_,S,Aerial,,15,"2,000x2,000",0.5,RGB
`Deep Globe Land Cover`_,S,DigitalGlobe +Vivid,803,7,"2,448x2,448",0.5,RGB
`DFC2022`_,S,Aerial,"3,981",15,"2,000x2,000",0.5,RGB
`ETCI2021 Flood Detection`_,S,Sentinel-1,"66,810",2,256x256,5--20,SAR
`EuroSAT`_,C,Sentinel-2,"27,000",10,64x64,10,MSI
`FAIR1M`_,OD,Gaofen/Google Earth,"15,000",37,"1,024x1,024",0.3--0.8,RGB
`Forest Damage`_,OD,Drone imagery,"1,543",4,"1,500x1,500",,RGB
Copy link
Collaborator Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Still don't know the resolution of this. It looks like Dan Morris (AI for Earth) is associated with this project and may know. There is also contact info on the dataset webpage: https://lila.science/datasets/forest-damages-larch-casebearer/

Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Sent an email to the author

Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

FYI Dan is at google now

@adamjstewart adamjstewart requested a review from calebrob6 July 2, 2022 23:38
calebrob6
calebrob6 requested changes Jul 5, 2022
View reviewed changes
Copy link
Member

@calebrob6 calebrob6 left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Can you change "mask" --> "masks"? I'll open an issue to update resolution on Forest Damages dataset.

calebrob6
calebrob6 previously approved these changes Jul 5, 2022
View reviewed changes
@calebrob6 calebrob6 closed this Jul 6, 2022
@calebrob6 calebrob6 reopened this Jul 6, 2022
@calebrob6 calebrob6 enabled auto-merge (squash) July 6, 2022 13:51
@calebrob6 calebrob6 merged commit 49d6114 into main Jul 7, 2022
@calebrob6 calebrob6 deleted the docs/tables branch July 7, 2022 05:11
@adamjstewart adamjstewart mentioned this pull request Jul 11, 2022
Merged
@adamjstewart adamjstewart added this to the 0.3.0 milestone Jul 11, 2022
yichiac pushed a commit to yichiac/torchgeo that referenced this pull request Apr 29, 2023
@adamjstewart
Improvements to dataset tables (microsoft#649)
7f5bba0 
* Improvements to dataset tables

* mask -> masks
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Reviewers

@calebrob6 calebrob6 calebrob6 approved these changes

Assignees
No one assigned
Labels
documentation Improvements or additions to documentation
Projects
None yet
Milestone
0.3.0
Development

Successfully merging this pull request may close these issues.
2 participants
@adamjstewart @calebrob6