Immobilized fluorescently stained zebrafish through the eXtended Field of view Light Field Microscope 2D-3D dataset
Authors/Creators
- 1. Computational Imaging and Inverse Problems, Department of Informatics, School of Computation, Information and Technology, Technical University of Munich, Germany 2Munich Institute of Biomedical Engineering, Technical University of Munich, Germany
- 2. Synthetic Neurobiology Group, Massachusetts Institute of Technology, USA
- 3. Computer Vision Group, University of Bern, Switzerland
Description
Immobilized fluorescently stained zebrafish through the eXtended Field of view Light Field Microscope 2D-3D dataset
This dataset comprises three immobilized fluorescently stained zebrafish imaged through the eXtended Field of view Light Field Microscope (XLFM, also known as Fourier Light Field Microscope). The images were preprocessed with the SLNet, which extracts the sparse signals from the images (a.k.a. the neural activity).
If you intend to use this with Pytorch, you can find a data loader and working source code to load and train networks here.
This dataset is part of the publication: Fast light-field 3D microscopy with out-of-distribution detection and adaptation through Conditional Normalizing Flows.
The fish present are:
- 1x NLS GCaMP6s
- 1x Pan-neuronal nuclear localized GCaMP6s Tg(HuC:H2B:GCaMP6s)
- 1x Soma localized GCaMP7f Tg(HuC:somaGCaMP7f)
The dataset is structured as follows::
XLFM_dataset
- Dataset/
- GCaMP6s_NLS_1/
- SLNet_preprocessed/
- XLFM_image/
- XLFM_image_stack.tif: tif stack of 600 preprocessed XLFM images.
- XLFM_stack/
- XLFM_stack_nnn.tif: 3D stack corresponding to frame nnn.
- Neural_activity_coordinates.csv: 3D coordinates of neurons found through the suite2p framework.
- XLFM_image/
- Raw/
- XLFM_image/
- XLFM_image_stack.tif: tif stack of 600 raw XLFM images.
- XLFM_image/
- SLNet_preprocessed/
- (other samples)
- GCaMP6s_NLS_1/
- lenslet_centers_python.txt: 2D coordinates of the lenset in the XLFM images.
- PSF_241depths_16bit.tif: 3D PSF of the microscope can be used for 3D deconvolution. Spanning 734 × 734 × 550𝜇𝑚3 used to deconvolve this volumes.
In this dataset, we provide a subset of the images and volumes.
Due to space constraints, we provide the 3D volumes only for:
- SLNet_preprocessed/XLFM_stack/
- 10 interleaved frames between frames 0-499 (can be used for training a network).
- 20 consecutive frames, 500-520 (can be used for testing).
- raw/
- No volumes are provided for raw data, but they can be reconstructed through 3D deconvolution.
Enjoy, and feel free to contact us for any information request, like the full PSF, 3 more samples or longer image sequences.