[Insight-developers] Image as an OrientedImage Progress

[Bill Lorensen]

Kent,

The Analyze tests use the NiftiImageIO it seems. Your changes to
AnalyzeImageIO are not fixing the errors in those tests.

Bill

On Tue, Sep 16, 2008 at 4:50 PM, Bill Lorensen <bill.lorensen at gmail.com> wrote:
> We need an N-D solution I think. Having a separate spatial dimension
> for coordinates and directions may be viable.
>
> Let's keep discussing and thinking...
>
> On Tue, Sep 16, 2008 at 3:42 PM, kent williams
> <norman-k-williams at uiowa.edu> wrote:
>> A more expedient question: Does it really make any sense to throw an
>> exception for degenerate 2D direction cosines. In those cases where valid 3D
>> dir cosines are truncated to degenerate 2D dir cosines, there's no valid
>> punt option, categorically:
>>
>> If you see dir cosines of [0 1] [0 0] or [0 0] [0 1] etc, what that means is
>> the second fastest moving direction is going in either the Z or -Z
>> direction.  You can cange them to e.g. [0 1] [1 0] or [0 1] [-1 0], which
>> will remove the degeneracy, but be 100% wrong.
>>
>>
>>
>> On 9/16/08 10:58 AM, "Bill Lorensen" <bill.lorensen at gmail.com> wrote:
>>
>>> As I said, although the issue has been known for some time, now is the
>>> time to attack it. We need some thought time on this one.
>>>
>>> For now, we'll just let the tests keep failing. One platform in the
>>> nightly is enough to keep us reminded.
>>>
>>> Bill
>>>
>>
>>
>