what is the reqiured preprocessing for new input images?

[aabdel-kader]

I'm trying to use the model with my own images but the resulting 3d mesh is always have problems with the z axis.
Is there any preprocessing required to be done for the input images before using the model in inference?

[bianwenjing]

As the model is trained using ShapeNet with white background, you will need to mask the image during inference.

[gokceay]

I'm trying to use the model with my own images but the resulting 3d mesh is always have problems with the z axis. Is there any preprocessing required to be done for the input images before using the model in inference?

Hi Aabdel, as far as I can understand you have created your own custom dataset. I am also trying to create my custom rayonet dataset but I am having error : 'Error occured when loading field points_view of model ...' during training. I think the problem is related with input images but I cannot find any error. Did you have a similar problem while generating your custom dataset? How did you generate your custom dataset?
Thanks in advance