Do a demo

[JadenLy]

I am interested in doing a demo on my own videos. But there is no model provided, so do I need to train the model by myself first on CDNet? Can you provide me with a trained model, since CDNet videos are too slow to download in China?

Besides, can you tell me how to make a call to test on my own videos. Your README does not provide any tutorial on how to use the files in this model.

Thanks!

[zhimingluo]

This work is not a general purpose foreground segmentation model, it mainly focus on the problem of labeling data, which takes some labeled frames from one video for training and the rest in the same video for testing.

[JadenLy]

@zhimingluo So the provided files will not generate an output using my own data and a trained model?