pretrained VAE provided by MAISI doesn't work well

[paproaskwar]

when i put 3D CT(.nii.gz) into the pretained VAE provided by MAISI, using autoencoder.reconstruct function, the reconstruct image looks different compared with the input one. it becomes wider

Can you tell me where the problem is
Very very appreciate!

[paproaskwar]

right is original drawing, and the left is the reconstruct image. the reconstruct image looks much "fatter" than the original, but they have a same shape(256, 256, 256)

[paproaskwar]

[paproaskwar]

I konw why...it seems that the output has a default space(1mm, 1mm, 1mm), which is different from input(1.5, 1.5, 2). Now im trying to find out how to change the output's space

[KumoLiu]

Please refer to https://github.com/Project-MONAI/tutorials/tree/main/generation/maisi#31-3d-autoencoder-training.