Luca Carini  0:05  
Hello. Did you have a slide?

Thank you Anne for making this slide, and if you do, click here, you'll see some use cases from us. Which, are you able to click through Anne? Don't worry if not don't worry if not, I'll describe, I can describe them. So I'm Luca from the Victoria and Albert Museum, where curators, conservators, researchers, and casual browsers have been comparing objects and their images since the museum's conception; and close comparison as always demanded some alignment of some sort. And finally in the eye of the observer. Oh brilliant thanks Anne, so you here you can see a list of roughly categorised use cases that we've encountered so far for image registration alignment. And there links to, each thing. You're welcome to try out later on. So working with images of objects, any misalignment, as a result of the imaging or scanning process is locked into the image for posterity. And when comparing images those misalignments have to be compensated for by the observer. Since adopting IIIF we've been using its image API to realign images on the fly. For close comparison in some of our image viewers, which you'd see if you click through any of those links. This alleviates the need for potentially endless image derivatives, depending on the comparisons being made, around any numbers specific regions of interest within an image. From my perspective as a developer of the V&A website. Some cases for image registration are simpler than others. In the case of our most recent digitization project, the job of registering the various scan types of the Raphael cartoons, fell to the team delivering us the scans as pre-registered images. However, since then we've also encountered our most complicated example for alignment in  a viewer, that of the cartoons, with the tapestries that once mirrored them precisely, but distorted in the centuries since they were made. We can now pull in the tapestry images directly from the Vatican Museum collections, but it's obvious we can only go so far at realigning them using the on the fly, crop, scale and rotate image API parameters, and as such we started wondering about the feasibility of extending the API's power to include warping or skewing, for example, but realise this would pose some big technical challenges, not least around processing power and caching implications, maybe a more practical approach would be to use the Presentation API to record and replay more complex image operations within IIIF viewers, that are able to process, extra image manipulation instructions, recorded within image manifests. These are some of the subjects, we'd like to discuss as our part of this webinar. And that's me, thanks.