Conversation
…ecure: just return the filepath, otherwise the file cant be proberly closed.
|
Let's move this to the |
|
There have been some conversations about whether to use a standard sklearn-like
|
Codecov Report
@@ Coverage Diff @@
## master #3267 +/- ##
==========================================
+ Coverage 86.81% 86.81% +<.01%
==========================================
Files 339 341 +2
Lines 27385 27403 +18
==========================================
+ Hits 23773 23790 +17
- Misses 3612 3613 +1
Continue to review full report at Codecov.
|
|
@emmanuelle @stefanv . I rebased the code and the tests are passing. |
|
@glemaitre thank you for the changes! I would support to have an API closer to scikit-learn with fit and predict, since users of this feature will probably have some background in machine learning (and scikit-learn), but I would think it's better to merge the PR and then do another PR changing the API for Cascade, ORB, ransac etc. |
| ========================================= | ||
|
|
||
| This example shows how to detect faces on an image using object detection | ||
| framework. |
There was a problem hiding this comment.
Maybe add a sentence like "It is an illustration of how to use machine learning for computer vision" ?
|
|
||
| plt.show() | ||
|
|
||
| """ |
There was a problem hiding this comment.
I don't think the following lines are required, since there is only one figure.
skimage/data/_detect.py
Outdated
| data_dir = _os.path.abspath(_os.path.dirname(__file__)) | ||
|
|
||
|
|
||
| def frontal_face_cascade_xml(): |
There was a problem hiding this comment.
The name is not very clear, I would prefer frontal_face_cascade_filename but it might be a matter of taste only, feel free to ignore this particular remark.
|
@emmanuelle I addressed the issues. |
|
The error is independent. The Travis build should be restarted. |
|
Thanks @glemaitre +1 for merging. @stefanv the ball is yours :-) |
|
It's very exciting to see this work finally make its way into skimage! Congratulations @warmspringwinds, and thank you very much @glemaitre for all the work you put in to complete this feature. |
|
I'll also add that we are all on board with trying to make a scikit-learn compatible interface for this and other feature detectors. |
|
That's amazing! Thank you so much @glemaitre for all your work and @vighneshbirodkar @stefanv @jni for mentoring during this project! Hopefully this new feature will find a lot of applications in the scikit-image community :) |
7 participants
superseded #1570
This PR is trying to rebase the original PR #1570