Dependencies

The code uses the following libraries

• NumPy
• SciPy
• Matplotlib
• Scikit-Learn
• ipyparallel for parallel processing
• opencv for efficient image manipulation and visualization
• Tifffile For reading tiff files. Other choices can work there too.
• cvxpy for solving optimization problems (for deconvolution, optional)
• Spams for online dictionary learning (for behavioral analysis, optional)

For the constrained deconvolution method (deconvolution.constrained_foopsi) various solvers can be used, some of which require additional packages:

1. 'cvxpy': (default) For this option, the following packages are needed:

• CVXOPT optional.
• CVXPY optional.

2. 'cvx': For this option, the following packages are needed:

• CVXOPT optional.
• PICOS optional.

In general 'cvxpy' can be faster, when using the 'ECOS' or 'SCS' solvers, which are included with the CVXPY installation. Note that these dependencies are circumvented by using the OASIS algoritm for deconvolution.