search

mehta-lab / recOrder

3D quantitative label-free imaging with phase and polarization
BSD 3-Clause "New" or "Revised" License
13 stars 4 forks source link

Denoising of fluorescence channels #80

Closed JohannaRahm closed 2 years ago

JohannaRahm commented 2 years ago

Issue: We want to reconstruct phase and denoise the channels simultaneously. Denoising of phase works. However, there is an issue when defining the fluorescence channel as target channel. It does not recognize the name of the channel, using index instead of str to define the channel does not work. The created zarr consists of black pixels only.

The current workaround is: 'So currently in the recOrder denoising, we can only denoise the hardcoded channels (S0-S3, BF, Retardance, Orientation, Phase2D, Phase3D). Fluorescence denoising is not implemented -- I would recommend using a notebook style reconstruction for this workflow and use the convenience functions in qlipp_compute.py and fluorescence_compute.py'

Command: recOrder.reconstruct --config /hpc/projects/comp_micro/projects/HEK/2022_01_20_orgs_BF_GFP_63x_04NA_JR/recOrder_issues/config_phase3D_denoising_phase.yml

Output data: /hpc/projects/comp_micro/projects/HEK/2022_01_20_orgs_BF_GFP_63x_04NA_JR/recOrder_issues/Phase3D_Denoising.zarr /hpc/projects/comp_micro/projects/HEK/2022_01_20_orgs_BF_GFP_63x_04NA_JR/recOrder_issues/Phase3D_Denoising_Flu.zarr

Branch: calibration_plugin Setup: running via CLI, HPC Data info: channel 0 = BF, channel 1 = GFP

Here the error message & the definition of denoising in the yml file:

image

ziw-liu commented 2 years ago

We have removed fluorescence-related features from the scope of recOrder.

@JohannaRahm, @talonchandler I think we can close this as well as #73 and #81, and remove them from the 1.0.0 Milestone if we have no plans to implement these.

talonchandler commented 2 years ago

Thanks for flagging @ziw-liu. Closing this.