Beautiful tool, but 2 some small "UI things"

[NormanTUD]

trafficstars

Hi, I've been looking for such a tool and at first got to CVAT. It looked nice, but my CPU does not support AVX2 and from there, troubles started. I tried 3 or 4 days installing it before giving up and looking around for other tools.

Yours worked out-of-the-box. Absolutely like I expected it, and there was nothing frustrating about this at all. So very good! And it does exactly what I want.

But now the 2 issues I think fixing would make the tool better in an instance. Luckily, they are small.

1. I don't see the sense of the "image overview" site you get when clicking on an image in a dataset and there you'll have to press the "Annotate set"-button. Just go to annotate directly. The screen is also misleading because you see coordinates under your mouse, implying they can do something here, but to the best of my knowledge as a very new user here, they can't.
2. Sometimes in firefox, when I try to draw a box and I try to click, it is registered as dragging the image.

For 2, please add the css-properties:

    user-drag: none;
    -webkit-user-drag: none;
    user-select: none;
    -moz-user-select: none;
    -webkit-user-select: none;
    -ms-user-select: none;

to the image. Imho this image should never be draggable (only via the hand tool).

[NormanTUD]

A third one may be:

Let people label everything they want and let them name it what they want, and, instead of prescribing a category for these labels, admins can then set that category X contains labels x, y and z.

This would be way more flexible for new datasets imho.

But still great tool, everything works in my tests.

[drewlr]

Hey @NormanTUD , thanks so much for the feedback, much appreciated. And really happy to hear Remo is serving you well :)

To answer to your Qs -

1. Image overview - the use case we have there is not for people that want to annotate a picture, rather for those that want to explore the annotations (or predictions) that a image has in details. That's also the thinking behind the coordinates - many times, in my experience, you just want to make sense of some annotation, and visualizing them along with coordinates is very useful

2- Yes that's not intended behaviour, thanks a lot for the CSS 👍 we are unfortunately not doing active development on Remo at the moment (and thinking what to do with it), so we won't be adding the code at this point in time. Hope the issue is not too annoying for you

3- That's a totally fair point, and it's something we had early on. What we noticed is that in few different use cases, an extra layer of governance is desirable - e.g. if you have an external team of annotators, and you don't want them to label things as they please. We tried to strike a balance in the design between extra governance and immediate of use