Re: [OctDev] imrotate

[Paul K. <pki...@us...>]

Yes, imrotate seems to be broken for all of bilinear, bicubic
and nearest.  As Todd pointed out, 'nearest' is particularly
glaring.

Can the users of this function please come up with a fix which
you are all happy with?

Here are some more tests you might include:

   % Make sure the dithering preserves pixel density
   t=zeros(1,5);
   for n=1:15; t(n) = sum(imrotate(eye(n),20,'bilinear')(:)); end
   assert(t,[1:15]);

   % Make sure the size matches the expected size assuming square pixels.
   angle=20; t=zeros(15,2);
   for n=1:15; t(n,:) = size(imrotate(ones(n,3*n),angle,'bilinear')); end
   a = angle*pi/180;
   x = ceil([(1:n);3*(1:n)]'*[cos(a) sin(a);sin(a) cos(a)]);
   assert(t,x);

Thanks,

- Paul

On Feb 19, 2005, at 2:55 PM, Jeff Orchard wrote:

>
> I'm not sure the bilinear option is working.
>
>   imrotate(eye(13),45,"bilinear")
>
> and
>
>   imrotate(eye(13),40,"bilinear")
>
> Jeff
>
>
> On Feb 16, 2005, at 12:48 AM, Todd Neal wrote:
>
>> On Tue, 15 Feb 2005 23:58:04 -0500, Paul Kienzle 
>> <pki...@us...> wrote:
>>> I replaced the old version of imrotate with what you sent.
>>>
>>> I would appreciate a few simple test cases which do not
>>> rely on a human visual system to determine if the result
>>> is correct.
>>>
>>> I tried the following:
>>>
>>>         X = rand(30);
>>>         Y = imrotate(imrotate(X,30,'bicubic'),-30,'bicubic');
>>>         norm(X(17+[1:5],17+[1:5])-Y(30+[1:5],30+[1:5]))
>>>         ans = 0.46103
>>>
>>> Using a smoother function worked better:
>>>
>>>         X = peaks(30);
>>>         Y = imrotate(imrotate(X,30,'bicubic'),-30,'bicubic');
>>>         norm(X(17+[1:5],17+[1:5])-Y(30+[1:5],30+[1:5]))
>>>         ans = 0.0036647
>>>
>>> Anyone have any better suggestions?
>>>
>>
>> I was looking for a few good test cases and came across using this as
>> a test case:
>>
>> imrotate(eye(20),45)
>>
>> The problem is displayed here (in a smaller case to fit in an email 
>> well):
>>
>> octave:18> imrotate(eye(13),45)
>> ans =
>>
>>   0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0
>>   0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0
>>   0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0
>>   0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0
>>   0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0
>>   0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0
>>   0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0
>>   0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0
>>   1  0  0  1  0  0  0  1  0  0  1  0  0  0  1  0  0  1
>>   1  0  0  1  0  0  0  1  0  0  1  0  0  0  1  0  0  1
>>   0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0
>>   0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0
>>   0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0
>>   0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0
>>   0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0
>>   0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0
>>   0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0
>>   0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0
>>
>>
>> Todd
>