Suggestion: Loading subset of image as defined by coordinate set only

Hello,

this is kind of related to #14, but only tangentially so.

Image-loading takes quite a bit of time if you are processing 4k x 6k sized images.

In fact it takes so prohibitively long that I am currently trying to figure out how to load subset of an image only. Imager does not allow for this; but RBioFormats does. However, setting up that package properly has turned out to be very complicated.

A good amount of the image is unlikely to be relevant for me, and I'd rather just not load it to begin with and save myself a lot of time and resources.

Could imager::load.image() be expanded by 4 parameters x1,x2,y1,y2 defining the range of pixel locations imported?

As a reference, the RBioFormats has a suitable function that does this in principal.

RBioFormats::read.image(
                                file = image.path,
                                ...,
                                subset = list(x=x1:x2,y1:y2)
                            )

Unfortunately, I have found that its image-data is not 1-to-1 with the data loaded by imager. I would not have assumed those to load identical data, but I certainly hoped for it. At this point I am not sure if the value differences are small enough so that I can deprecate imager::load.image() for RBioFormats::read.image() to implement this; because I have not come around to testing this particular change.

Nonetheless I would prefer a native support in imager so that I don't have to worry about these value diffs. Additionally, setting up RBioFormats() is somewhat of a pain to manage; particularly when you must consider it in the scope of a package, instead of a "normal" script-environment.

Is something like this possible in principal?

Thank you. Sincerely, ~Gw

asgr / imager

Suggestion: Loading subset of image as defined by coordinate set only #15