Describe the bug
I have a script for running Cellpose and after adding options to use the DenoiseModel or CellposeDenoiseModel, the output for a 3D image is flattened across one of the axes when using the denoising related models. I've confirmed that the tiff file is being loaded properly and the default CellposeModel is acting as expected, returning a 3D image. I'm thinking there's a setting I am getting wrong or missing, but I'm unable to find a solution so far.
Run log
INFO:cellpose.io:reading tiff with 350 planes
100%|██████████| 350/350 [00:10<00:00, 34.12it/s]
INFO:cellpose.denoise:>> denoise_cyto3 << model set to be used
INFO:cellpose.core: TORCH CUDA version installed and working.
INFO:cellpose.core:>>>> using GPU
INFO:cellpose.denoise:>>>> model diam_mean = 30.000 (ROIs rescaled to this size during training)
Scan_Iter_0003_CamA_ch0_CAM1_stack0000_488nm_0000000msec_0018536711msecAbs_decon.tif has shape (350, 1480, 1879)
Denoised image has shape (1, 350, 1480, 1)
Describe the bug I have a script for running Cellpose and after adding options to use the DenoiseModel or CellposeDenoiseModel, the output for a 3D image is flattened across one of the axes when using the denoising related models. I've confirmed that the tiff file is being loaded properly and the default CellposeModel is acting as expected, returning a 3D image. I'm thinking there's a setting I am getting wrong or missing, but I'm unable to find a solution so far.
Run log INFO:cellpose.io:reading tiff with 350 planes 100%|██████████| 350/350 [00:10<00:00, 34.12it/s] INFO:cellpose.denoise:>> denoise_cyto3 << model set to be used INFO:cellpose.core: TORCH CUDA version installed and working. INFO:cellpose.core:>>>> using GPU INFO:cellpose.denoise:>>>> model diam_mean = 30.000 (ROIs rescaled to this size during training) Scan_Iter_0003_CamA_ch0_CAM1_stack0000_488nm_0000000msec_0018536711msecAbs_decon.tif has shape (350, 1480, 1879) Denoised image has shape (1, 350, 1480, 1)