MinMaxLTTB: Leveraging MinMax-Preselection to Scale LTTB

Abstract

Figure 1: Visual analysis of MinMax (column 1), MinMaxLTTB with preselection ratio r ps ∈ {2, 4} (column 2-3), and LTTB (column 4). Each row corresponds to a unique template, with the rightmost column presenting the reference images, constructed from 200k data points. The four adjacent aggregations utilize an equal number of output samples (n out = 200). Columns 2 and 3 illustrate the effect of the preselection ratio r ps for MinMaxLTTB. Visualization properties used include: Plotly as toolkit, line-width of 2 pixels, and anti-aliasing enabled. A GIF further demonstrates the above visualization for varying values for n out .