Feature Request: FOMO Grid

Just ignore me if I am doing too many feature requests:

From the notes here I read:

"In short, if we make the following two constraining assumptions:

1. All bounding boxes are square and have a fixed size
2. The objects occupy a grid over the input

Then we can vastly reduce the complexity, and hence the size and speed of our model. In that case, FOMO will work in its optimal condition."

I suggest for setting FOMO bounding boxes, that we have a grid over the image and just select grid squares for the objects. Bonus: the ability to shift the image up, down, left, right to better line up the objects inside the grid before selecting the objects grid locations.

Actually what I want is just click on objects to label them, rather than bounding boxes. This would also work once we ship different factors (8:1 now, but want to do 2:1 too).

Yes, clicking objects would be great, (also fewer alert popups the better) . I am looking forward to any heat map smaller than 8:1 even if it slows my Portenta 96x96, 72ms analysis a bit.

6:1 or 4:1 would be fine. I would love to test it.

I just tried the FOMO Queue. It was great, at first I didn’t realize what it was doing, since my photos were very similar.

The Queue is amazing, everyone, make sure you try it.

1 Like