Skip to content

test data for "New camera model for stitching pipeline" #6933#303

Merged
alalek merged 1 commit intoopencv:masterfrom
hrnr:gsoc_all
Oct 22, 2016
Merged

test data for "New camera model for stitching pipeline" #6933#303
alalek merged 1 commit intoopencv:masterfrom
hrnr:gsoc_all

Conversation

@hrnr
Copy link
Copy Markdown
Contributor

@hrnr hrnr commented Jul 17, 2016

this is for new perf test in calib3d: regression data for EstimateAffine test

opencv/opencv#6933

@hrnr hrnr changed the title regression data for EstimateAffine test #6933 test data for "New camera model for stitching pipeline" #6933 Aug 11, 2016
@alalek
Copy link
Copy Markdown
Member

alalek commented Oct 12, 2016

@hrnr
Could you squash all commits into one (this will remove intermediate changes and should reduce size of patch)?

squashed

* perf: add regression data for EstimateAffine test

* add test images for affine stitching tests
created from https://commons.wikimedia.org/wiki/File:Pont_du_Gard_BLS.jpg splitted and resized, so there is only a translation between both parts

* update regression data for rewritten stitching tests

stitching tests has been rewriten to check sanity on estimated transformation. Values should be now comparable for surf and ORB.

* add perf data for affine stitching tests

For sanity check at match_affineBestOf2Nearest. Trasformation is comparable between surf and orb

* add sanity data for affine warper ocl perf test

* stitching: add perf sanity for affine bundle adjuster test

* stitching: update perf data for estiamtors test

test has been rewritten to test both affine and affinePartial bundle adjusters

* add new data for affine stitching tests

images data from scanners. 'budapest' dataset is from professional book scanner Minolta PS 7000. It is a scan of foldable map, therefore not perfectly flat. 'newspaper' and 'prague' datasets are from home-grade color scanner HP Photosmart C3100 All-in-One.

All images has been resized with imagemagic, original size ~600MB.

* remove sanity data for calib3d test

replaced by accuracy test, since the amount of test cases is now much more bigger
@hrnr
Copy link
Copy Markdown
Contributor Author

hrnr commented Oct 12, 2016

squashed, but patch size is not much smaller. The main problem are images that are needed for tests. They are already resized and compressed.

Let me know if I need to reduce the size more.

@alalek
Copy link
Copy Markdown
Member

alalek commented Oct 12, 2016

Thanks! Looks good to me 👍

Could you please resolve merge conflict in the main PR?

@alalek alalek merged commit 915365a into opencv:master Oct 22, 2016
tucna pushed a commit to tucna/opencv_extra that referenced this pull request Dec 3, 2016
test data for "New camera model for stitching pipeline"

* perf: add regression data for EstimateAffine test

* add test images for affine stitching tests
created from https://commons.wikimedia.org/wiki/File:Pont_du_Gard_BLS.jpg splitted and resized, so there is only a translation between both parts

* update regression data for rewritten stitching tests

stitching tests has been rewriten to check sanity on estimated transformation. Values should be now comparable for surf and ORB.

* add perf data for affine stitching tests

For sanity check at match_affineBestOf2Nearest. Trasformation is comparable between surf and orb

* add sanity data for affine warper ocl perf test

* stitching: add perf sanity for affine bundle adjuster test

* stitching: update perf data for estiamtors test

test has been rewritten to test both affine and affinePartial bundle adjusters

* add new data for affine stitching tests

images data from scanners. 'budapest' dataset is from professional book scanner Minolta PS 7000. It is a scan of foldable map, therefore not perfectly flat. 'newspaper' and 'prague' datasets are from home-grade color scanner HP Photosmart C3100 All-in-One.

All images has been resized with imagemagic, original size ~600MB.

* remove sanity data for calib3d test

replaced by accuracy test, since the amount of test cases is now much more bigger
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants