Welcome to fourS

This is the documentation of fours, a Python package for PSF subtraction with the 4S algorithm for exoplanet high contrast imaging (HCI).

In this documentation, we explain the how to use the package and provide some guides to help you get started. We also provide the code to reproduce the plots in our paper. fours is fully compatible with other HCI packages like VIP or PynPoint. Detection limits can be computed with applefy.

Please read the section about Citation if you use fours.

Precovery of AF Lep b

Using the 4S algorithm, we were able to recover the planet AF Lep b in archival data from 2011. In the following gallery we show the 4S and PCA residuals for this dataset.

The planet is clearly visible in the 4S residuals irrespective of the regularization parameter (lambda). Given the knowledge of the planet position, we can also see it in the PCA residuals (set the components to 65 and combine to mean).

Finding new planets with 4S

In the following gallery we show residual images obtained with 4S and the widely used PCA algorithm. 3 Fake planets were injected into the data.

Both algorithms have one free hyper-parameter. For PCA this is the number of components used for the PSF model. For 4S this is the regularization parameter lambda.

Note: In the PCA residuals the signal of the fake planets sometimes co-aligns with speckles. In practice, this can lead to a biased estimate of the planet photometry! You can switch between the examples to see if the current PCA residual is affected.