Prediction of amodal mask for own labeled dataset

[zhangyufeng1998]

hi,thanks for youe work! I labeled the image with the labeling tool as shown Why does it produce a modal mask and an amodal mask that are almost identical? Why does it produce 3 boxes？

[kkllww]

Hello, can you download the released model.

[jiunyen-ching]

Hi @zhangyufeng1998, in your annotation screenshot, we can see that you indeed labelled the segments of two leaves correctly.

However, in the visualization step, we can see that there are INPUT 3 modal masks (highlighted by 3 different colors)

This means that you need to modify the code that reads your annotations to load the masks correctly before you run de-occlusion. Verify that the modal masks are being visualized as how you would expect the masks to be.

[zhangyufeng1998]

ok i have got it now,i open the evaluation issue ,can you help me ? It is the issues 50th

[wmrenr]

Hello, can you download the released model. Yes, I can download with vpn tool.

[fserk729]

hello @zhangyufeng1998

how did you make your own dataset for this?