Figure 2.
(A) Schematic of the algorithm and the optimization parameters. (B) Effect of the maximum box size in the box-counting algorithm on the R 2 value of the linear regression of log of box sizes and boundary-containing box counts. The maximum box size was defined as the number of pixels of the x-axis of the image divided by the box size parameter and rounded down to the nearest integer. (C) Optimization results based on the difference between the final fractal dimension estimate for LEC−LEC and BEC−BEC condition and LEC−BEC condition. (D) Optimization result based on the average standard error between the estimated fractality of each timeseries image of the same condition.