Handling big number of annotations

[mcondor10]

Hi! I am working in cell instance segmentation, in which a single image can have thousands of labels. I already have a trained model which helps creating these labels, but sometimes they require some corrections. I was wondering if there are some guidelines or if I can have some advice about how to handle that number of annotations. Initially, I had them as masks, but noticing how slow it made the browser, I changed to polygons. This helped but the issue is still present.

Any advice would be appreciated, and thanks for the great project! :)

[makseq]

I can advise only one thing - try to hide all regions by default: https://labelstud.io/guide/predictions.html#Read-only-and-hidden-regions