Closed huangjunxiong11 closed 4 years ago
The principle of the paper requires us to provide two pictures:images and the image's trimap
Get Trimap's code in here: https://github.com/foamliu/Deep-Image-Matting-PyTorch/blob/db6a96de7dd28e0a6746ec4d9a1c2efdf200c701/data_gen.py#L87
But this function need we provide alpha image. It seems we have to get it by ourself. Obtained manually by PS or by semantic segmentation model
This is my personal view,I don't know if this is right.
你好,请问你是否能提供预训练模型以供参考?
It depends on trimaps.
Now, I got a model, I want to use it but I can't, because I have not the Trimaps of my pictures. Are there the script of code to build the Trimaps? How can i get the Trimaps of my pictures?