Tuning the hyper-parameter μ for the total-variation regularization with MiRA/SPARCO . Left: cost functions vs. the hyper-parameter μ. Red line: χ² (=). Black dashed line: (multiplied by 1000 for better graph visibility). Black solid line: . Blue line: . The vertical dotted line corresponds to the best regularization weight μ⁺ found by the quality criterium. Center: The L curve is one of the criteria to choose the regularization weight μ. Black line: the values of vs. for a range of μ values from 1 to 10⁶. Blue triangles: points where μ = {1,10,10²,10³,10⁴,10⁵,10⁶ }. Red triangle: the value where μ = μ⁺. Right: the quality criterium graph is the distance between the pixels of the reconstructed image to the real (here the model) one. Blue line: the value of the quality criterium vs. the hyper-parameter μ. Dotted line: the position of the minimum of the curve and the corresponding value of μ = μ⁺.