Issue with nonlinearregistration

[Vanwalleghem]

Hi,

we are using the protocol outlined in Brain Registration and Evaluation for Zebrafish (https://www.sciencedirect.com/science/article/pii/S2666166723006147). However, we are running into issues when using our full resolution images.

+ +  Preprocessing Step1 linear registration:level_ref_image:
+ +      Extracting frame =0  dim=(621,840,138) spa=(0.798,0.798,2.000) rng=(0.000:4342.006)
+ +      Smoothing (2.033,2.033,2.033) --> dim=(621,840,138) spa=(0.798,0.798,2.000) rng=(0.000:4093.156)  vx-sigmas=(2.548,2.548,1.017)
+ +      Resampling (4.788,4.788,4.788) -> dim=(104,140,58) spa=(4.788,4.788,4.788) rng=(0.000:4077.788)
+ +      Normalizing (0.010:0.990) 63 --> dim=(104,140,58) spa=(4.788,4.788,4.788) rng=(0.000:63.000)  robust 1:99 %, info=0.000,1695.080 numbins=64
+ +
+ +  Preprocessing Step1 linear registration:level_targ_image:
+ +      Extracting frame =0  dim=(1024,1024,127) spa=(0.692,0.692,1.500) rng=(0.000:65535.000)

The code works well with downsampled images, so could it be a memory issue?
I have tried increasing the node js max memory size with NODE_OPTIONS, but that doesn't change anything.

I tried running the code with debug on, but there was no more information, any advice would be welcome