Computer Vision News - July 2018

In this week’s section we will demonstrate the CVPXY library . As a bonus, we will show how you can run this on Colaboratory , a new Google environment for running python online that requires no setup to use. CVXPY is a dedicated Python toolbox aimed at solving convex optimization. Its user-friendly coding style inspired by CVX (Grant and Boyd, 2014) allows the user to express a wide range of convex optimization problems, in mathematically intuitive syntax. CVXPY is open source with GPL license available at http://www.cvxpy.org/ One programming approach for convex optimization is to use a Domain Specific Language (DSL). In this approach you get to express your problem in a mathematically intuitive way, and then it is automatically converted into the format needed by the solver (e.g., CVX (Grant and Boyd, 2014), YALMIP (Lofberg, 2004), QCML (Chu et al., 2013), PICOS (Sagnol, 2015), and Convex.jl (Udell et al., 2014)). CVXPY offers a new approach to DSL: on the one hand it’s a dedicated convex optimization DSL, and at the same time it’s an ordinary Python library, which allows you to take advantage of all the features of Python. In this article we will demonstrate the use of the CVXPY library to solve 4 classic problems : 1. Least-squares problem with box constraints 2. Lasso with Python built in multi-threading 3. Shortest path 4. Image in-painting 2 We Tried For You: CVXPY and Colaboratory Tool by Assaf Spanier CVXPY is a dedicated Python toolbox aimed at solving convex optimization. Its user- friendly coding style... Computer Vision News