[SciPy-Dev] imresize does not work with size = int

[Vincent Davis]

On Sat, Jul 3, 2010 at 5:29 PM, Vincent Davis <vincent at vincentdavis.net> wrote:
> I might be doing something wrong but..
> This is after the recent removal of the duplicate imresize
> proposed fix at bottom.
>
>>>> imresize(x, 2, mode='F')
> Traceback (most recent call last):
>  File "<stdin>", line 1, in <module>
>  File "/Library/Frameworks/Python.framework/Versions/2.7/lib/python2.7/site-packages/scipy/misc/pilutil.py",
> line 313, in imresize
>    imnew = im.resize(size, resample=func[interp])
>  File "/Library/Frameworks/Python.framework/Versions/2.7/lib/python2.7/site-packages/PIL/Image.py",
> line 1302, in resize
>    im = self.im.resize(size, resample)
> TypeError: must be 2-item sequence, not float
>
> THIS WORKS
>
>>>> imresize(x, 2., mode='F')
> array([[-0.654, -0.654, -0.227,  0.201,  0.141,  0.081],
>       [-0.654, -0.654, -0.227,  0.201,  0.141,  0.081],
>       [ 0.242,  0.242, -0.254, -0.751, -0.806, -0.86 ],
>       [ 1.139,  1.139, -0.282, -1.703, -1.752, -1.801],
>       [ 0.216,  0.216, -0.818, -1.851, -1.591, -1.332],
>       [-0.708, -0.708, -1.354, -1.999, -1.431, -0.862]], dtype=float32)
>
> AND THIS
>
>>>> imresize(x, (100,100), mode='F')
> array([[-0.654, -0.654, -0.654, ...,  0.081,  0.081,  0.081],
>       [-0.654, -0.654, -0.654, ...,  0.081,  0.081,  0.081],
>       [-0.654, -0.654, -0.654, ...,  0.081,  0.081,  0.081],
>       ...,
>       [-0.708, -0.708, -0.708, ..., -0.862, -0.862, -0.862],
>       [-0.708, -0.708, -0.708, ..., -0.862, -0.862, -0.862],
>       [-0.708, -0.708, -0.708, ..., -0.862, -0.862, -0.862]], dtype=float32)
>
> I think the problem is in line 307
> size = size / 100.0
> I think size needs to be a tuple
> looking at line 309 which is what is used for size=float
> size = (array(im.size)*size).astype(int)
>
> I think the fix is
> percent = size / 100.0
> size = (array(im.size)*percent).astype(int)

Applying that change seems to fix it. The following two operation
should get the same results and do.

>>> imresize(x, 200, mode = 'F')
array([[ 0.68348145,  0.68348145,  1.43379724,  2.18411303,  1.42521787,
         0.66632271],
       [ 0.68348145,  0.68348145,  1.43379724,  2.18411303,  1.42521787,
         0.66632271],
       [-0.41771227, -0.41771227, -0.07694912,  0.26381403,  0.07136334,
        -0.12108734],
       [-1.518906  , -1.518906  , -1.58769548, -1.65648496, -1.28249121,
        -0.90849739],
       [-0.43312848, -0.43312848, -0.3485043 , -0.26388013, -0.3611995 ,
        -0.45851886],
       [ 0.65264904,  0.65264904,  0.89068687,  1.12872469,  0.56009215,
        -0.00854035]], dtype=float32)
>>> imresize(x, 2., mode = 'F')
array([[ 0.68348145,  0.68348145,  1.43379724,  2.18411303,  1.42521787,
         0.66632271],
       [ 0.68348145,  0.68348145,  1.43379724,  2.18411303,  1.42521787,
         0.66632271],
       [-0.41771227, -0.41771227, -0.07694912,  0.26381403,  0.07136334,
        -0.12108734],
       [-1.518906  , -1.518906  , -1.58769548, -1.65648496, -1.28249121,
        -0.90849739],
       [-0.43312848, -0.43312848, -0.3485043 , -0.26388013, -0.3611995 ,
        -0.45851886],
       [ 0.65264904,  0.65264904,  0.89068687,  1.12872469,  0.56009215,
        -0.00854035]], dtype=float32)
>
> Vincent

>