PERF: ImageMaskSpatialObject use image regions from its base class

A significant performance improvement appears to be achieved by using the
buffered region and the requested region as cached in its base class, instead
of calling the virtual functions `GetBufferedRegion()` and
`GetRequestedRegion()` from its image.

A performance improvement of more than 8 percent was observed, from more than
0.80 sec. (before this commit) to less than 0.73 sec. (after this commit), when
calling `ImageMaskSpatialObject::IsInsideInObjectSpace` 2^27 times, using a
Visual Studio 2022 Release build.

Modules/Core/SpatialObjects/include/itkImageMaskSpatialObject.hxx

@@ -39,7 +39,7 @@ ImageMaskSpatialObject<TDimension, TPixel>::IsInsideInObjectSpace(const PointTyp
 
   const IndexType index = image->TransformPhysicalPointToIndex(point);
 
-  return image->GetBufferedRegion().IsInside(index) &&
+  return Superclass::GetBufferedRegion().IsInside(index) &&
          Math::NotExactlyEquals(image->GetPixel(index), NumericTraits<PixelType>::ZeroValue());
 }
 
@@ -163,7 +163,7 @@ ImageMaskSpatialObject<TDimension, TPixel>::ComputeMyBoundingBoxInIndexSpace() c
     return RegionType{ minIndex, regionSize };
   };
 
-  const RegionType requestedRegion = image.GetRequestedRegion();
+  const RegionType requestedRegion = Superclass::GetRequestedRegion();
 
   if (requestedRegion.GetNumberOfPixels() == 0)
   {