[ptx] Re: Enblend suggestion: focus score weighting

[Pablo d'Angelo]

On Thu, 17 Jun 2004, Andrew C Mihal wrote:

> On Wed, 16 Jun 2004, Sebastian Nowozin wrote:
> 
> Another interesting metric would be: this pixel is too far out of 
> consensus with other overlapping input pixels. This might be a way to 
> identify ghosts. You have to decide whether to keep the ghost or delete it 
> though.

I have already thought about something like that, maybe correlating small
area patches and using only one patch if the correlation coefficient is too
low.



Something else: I have worked on vigra's import/export module to support
arbitrary image types with alpha channel, and the hugin cvs contains
functions to to this, in case you're interested in that.

Currently there is a small bug somewhere that incorrectly converts the 16/32
bit or float alpha channels to 8 bit, but I'll fix that soon. The interface
should will stay the same and looks like that:

vigra_impex2::ImageImportInfo info(filename);
// total number of channels
int bands = info.numBands();
// number of extra (non color) channels, usually just one alpha mask,
// but could be more for tiff files.
int extraBands = info.numExtraBands();

// single channel image
if (bands == 1 || bands == 2 && extraBands == 1) {
    if (strcmp(pixelType, "UINT8") == 0 ) {
    	// 8 bit image
	BImage gray;
	BImage mask;
	// will add an all 255 alpha channel if image doesn't contain one.
	// usually one wants to call a template function that does all the
	// stitching/blending, instead of just importing the image.
        vigra_impex2::importImageAlpha(info, destImage(gray), destImage(alpha);
   } if (strcmp(pixelType, "FLOAT") == 0 ) {
   	FImage gray;
	BImage mask;
        vigra_impex2::importImageAlpha(info, destImage(gray), destImage(alpha);
   }
   
   ...
} else if (bands == 3 || bands == 4 && extraBands == 1) {

   ... 
   
}


Exporting images with alpha channel can be done with:
exportImageAlpha(srcImageRange(gray), srcImage(alpha), info);

for TIFF, 8, 16 and 32 bit images, as well as float and double images
are supported. PNG supports only uint8 and uint16 images.

ciao
  Pablo