Published November 5, 2020 | Version v1
Dataset Open

PSF Estimation and deconvolution: models, microscopy images, and datasets

  • 1. Idiap Research Institute

Description

This is the accompanying dataset for the publication Adrian Shajkofci, Michael Liebling, “Spatially-Variant CNN-Based Point Spread Function Estimation for Blind Deconvolution and Depth Estimation in Optical Microscopy,” IEEE Transactions on Image Processing, vol. 29, pp. 5848-5861, 2020.

Publications based on this data must cite the above paper.

BibTeX Citation:
@ARTICLE{shajkofci.liebling:20,
  author={A. Shajkofci and M. Liebling},
  journal={IEEE Trans. Image Proces.}, 
  title={Spatially-Variant {CNN}-Based Point Spread Function Estimation for Blind Deconvolution and Depth Estimation in Optical Microscopy},
  year={2020},
  volume={29},
  number={},
  pages={5848-5861},
  doi={10.1109/TIP.2020.2986880}}
 

In the archive, you will find :

  • - Trained models for PSF estimation and deconvolution
  • - Synthetic training dataset of cells and beads
  • - Stacks of multi-channel fluorescence microscopy images of HeLa cells, rat brain cells, beads and plant cells to test the PSF estimation tool, deconvolution algorithm or auto-focus algorithm.
  • - Stacks of tilted grid (3, 6 and 9 degrees) using astigmatic lenses for depth estimation.
     

The code for running the models is available here:
https://github.com/idiap/psfestimation


Reference paper

A. Shajkofci and M. Liebling, "Spatially-Variant CNN-Based Point Spread Function Estimation for Blind Deconvolution and Depth Estimation in Optical Microscopy," in IEEE Transactions on Image Processing, vol. 29, pp. 5848-5861, 2020, doi: 10.1109/TIP.2020.2986880.

 

Ethical compliance

The post-mortem stained and fixed tissue slices whose images are included in this data set were reused from experiments approved by the EPFL ethics committee.

 

Funding

This work was supported by the Swiss National Science Foundation under Grants 206021_164022 “Platform for Reproducible Acquisition, Processing, and Sharing of Dynamic, Multi-Modal Data” and 200020_179217 “COMPBIO: Computational biomicroscopy: advanced image processing methods to quantify live biological systems”

Files

MD5SUM.txt

Files (24.9 GB)

Name Size Download all
md5:9f7a33c6a785bbb905c045fdcb14b15d
260 Bytes Preview Download
md5:47b75c823a3025ce8c696e9ab86bbf36
2.0 GB Download
md5:8835f0ac340f80e0418cd7756cc17bd9
15.7 GB Download
md5:4fdeec04cb45ff55b238851d350021d4
508.6 MB Download
md5:55001d734edf26c48bf4724ab59b7595
6.7 GB Download
md5:37a8c508944d7bdd646ca863bf55bdca
1.6 kB Preview Download

Additional details

Related works

Is documented by
Journal article: 10.1109/TIP.2020.2986880 (DOI)

Funding

Platform for Reproducible Acquisition, Processing, and Sharing of Dynamic, Multi-Modal Data 206021_164022
Swiss National Science Foundation
Computational biomicroscopy: advanced image processing methods to quantify live biological systems 200020_179217
Swiss National Science Foundation