Rex Cheng

Does this help: https://github.com/hkchengrex/XMem/blob/main/docs/INFERENCE.md?

Did you see the "On custom data" section?

The Colab demo (in the readme, https://colab.research.google.com/drive/1RXK5QsUo2-CnOiy5AOSjoZggPVHOPh1m?usp=sharing) can probably help you

Might work better if you zoom into the tree region since there are different semantics in this scene. You can also try making trees the foreground. Re-training might help but...

What's the input image/mask?

There is an example command in the readme file. Does that not work for you?

@DUT-CSJ Are the precomputed results in the 0/1 format or 0~255 format as mentioned in the readme?

By default, the training goes from stage 0->3. The provided model is in stage 3, and the existing code does not support "going back" to stage 0. Basically the first...

If I understand you correctly, you can just not load the pretrained model.

I don't know what your data looks like but I think you need a strong temporal smoothness prior to segment clouds which our model does not provide.