OpenCV Python bindings for cv::ml::SVM::trainAuto

Question

I want to estimate the optimal C and gamma parameters for my SVM training in OpenCV. If I understand the master (3.1-dev) docs correctly cv::ml::SVM::trainAuto would be fit perfect for my needs (using cross-validation to estimate the perfect parameters).

But I can't find the Python Bindings for trainAuto.. I tried to find it using:

>>> import cv2
>>> cv2.__version__
'3.1.0-dev'
>>> help(cv2.ml)

or in the ml_SVM object

>>> help(cv2.ml.SVM_create())

But I only found

SVM_create(...)
    SVM_create() -> retval

SVM_load(...)
    SVM_load(filepath) -> retval

in the cv2.ml module and

train(...)
 |      train(trainData[, flags]) -> retval  or  train(samples, layout, responses) -> retval

in the ml_SVM object. Is there another "python way" for trainAuto or are the bindings moved/missing? I'm using python 3.4 on Ubuntu 15.10.

score -1 · Answer 1 · edited Nov 25 '16 at 16:43

-1

It is a current open issue with OpenCV for Python: see here https://github.com/opencv/opencv/issues/7224.

edited Nov 25 '16 at 16:43

Tunaki

132,869
46
340
423

answered Nov 25 '16 at 16:41

Ben Hazelwood

75
1
2
7

I know.. I opened the issue ... as you could see in the reference to SO ... "[...] Yesterday I wrote ---->this question<---- on stackoverflow [...]." – Fabian Nov 25 '16 at 17:33
Ah ok saw this a while ago and didn't make the link sorry – Ben Hazelwood Dec 02 '16 at 14:51

OpenCV Python bindings for cv::ml::SVM::trainAuto

1 Answers1