We’re tagging pairs of images (for training a siamese network model).
Is there a way to present two images at each tagged datapoint?
Right now we are just merging them into a single larger image, but wondering if there is a more elegant way to do it.
Hi! I think generating a new image is a fine solution. Alternatively, you could also use the HTML interface /
"html" property to create a layout with two images side-by-side. That might be a bit more predictable to generate programmatically, and you can reuse the original images and add more meta information if needed. Even if you’re using the base64-encoded strings, you can still include those –
<img src="data:... />" is totally valid.