From stefan at sun.ac.za  Thu Dec  2 10:13:21 2010
From: stefan at sun.ac.za (=?ISO-8859-1?Q?St=E9fan_van_der_Walt?=)
Date: Thu, 2 Dec 2010 17:13:21 +0200
Subject: OpenCV: Upgrade to 2.1 and addition of cvFindFundamentalMat
Message-ID: <AANLkTimw9=und_e3wtdFs8VVE+=XQMsgQZwK-KwM+sa2@mail.gmail.com>

Hi everyone,

I'd be glad if you could review the following changes.  First, a minor
upgrade of the OpenCV wrappers to bring us up to date with OpenCV 2.1:

https://github.com/stefanv/scikits.image/compare/master...OpenCV_2.1

Then, the addition of cvFindFundamentalMat:

https://github.com/stefanv/scikits.image/compare/OpenCV_2.1...cvFindFundamentalMat

Further, are there any Christmas wishes that we should get working on
before the end of the year?

Regards
St?fan


From sccolbert at gmail.com  Thu Dec  2 19:44:05 2010
From: sccolbert at gmail.com (Chris Colbert)
Date: Thu, 2 Dec 2010 18:44:05 -0600
Subject: OpenCV: Upgrade to 2.1 and addition of cvFindFundamentalMat
In-Reply-To: <AANLkTimw9=und_e3wtdFs8VVE+=XQMsgQZwK-KwM+sa2@mail.gmail.com>
References: <AANLkTimw9=und_e3wtdFs8VVE+=XQMsgQZwK-KwM+sa2@mail.gmail.com>
Message-ID: <AANLkTine4X78Uf1zdxYddPhMp7XPSakN75X3BAipYiw=@mail.gmail.com>

For the cvUndistort2, you could change line 2194 to call assert_same_shape()
which lives in opencv.opencv_backend.pyx
Ditto for line 2541 in cvFindFundamentalMat.

Could you elaborate on this comment:?
+    ## The code snippet below creates the ``status`` matrix
+    ## that may optionally be provided.  I could not get this
+    ## to work with OpenCV 2.1.

Cheers,

Chris


2010/12/2 St?fan van der Walt <stefan at sun.ac.za>

> Hi everyone,
>
> I'd be glad if you could review the following changes.  First, a minor
> upgrade of the OpenCV wrappers to bring us up to date with OpenCV 2.1:
>
> https://github.com/stefanv/scikits.image/compare/master...OpenCV_2.1
>
> Then, the addition of cvFindFundamentalMat:
>
>
> https://github.com/stefanv/scikits.image/compare/OpenCV_2.1...cvFindFundamentalMat
>
> Further, are there any Christmas wishes that we should get working on
> before the end of the year?
>
> Regards
> St?fan
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/scikit-image/attachments/20101202/2aa85ce1/attachment.html>

From stefan at sun.ac.za  Fri Dec  3 03:23:03 2010
From: stefan at sun.ac.za (=?ISO-8859-1?Q?St=E9fan_van_der_Walt?=)
Date: Fri, 3 Dec 2010 10:23:03 +0200
Subject: OpenCV: Upgrade to 2.1 and addition of cvFindFundamentalMat
In-Reply-To: <AANLkTine4X78Uf1zdxYddPhMp7XPSakN75X3BAipYiw=@mail.gmail.com>
References: <AANLkTimw9=und_e3wtdFs8VVE+=XQMsgQZwK-KwM+sa2@mail.gmail.com> <AANLkTine4X78Uf1zdxYddPhMp7XPSakN75X3BAipYiw=@mail.gmail.com>
Message-ID: <AANLkTi=VdFb6mSe+jH=3v7Fd1qZAwDdSygWeAufKAfic@mail.gmail.com>

On Fri, Dec 3, 2010 at 2:44 AM, Chris Colbert <sccolbert at gmail.com> wrote:
> For the cvUndistort2, you could change line 2194 to call assert_same_shape()
> which lives in opencv.opencv_backend.pyx

Looks like the two arrays are Nx3 and Nx2 respectively, so I just
checked the lengths.

> Ditto for line 2541 in cvFindFundamentalMat.
> Could you elaborate on this comment:?
> + ? ?## The code snippet below creates the ``status`` matrix
> + ? ?## that may optionally be provided. ?I could not get this
> + ? ?## to work with OpenCV 2.1.

I construct the `status` array and send it as the last parameter to
cvFindFundamentalMat.  It passes the internal shape check, but fails
*another* shape check later on.  I think this might be a bug in
OpenCV, or maybe there is something I don't understand.  If you just
uncomment the code snippet, and change the NULL to accept cvstatus,
you'll see what I mean.

Regards
St?fan