Skip to main content
Figure 3 | BioMedical Engineering OnLine

Figure 3

From: Algorithm for automatic analysis of electro-oculographic data

Figure 3

Baseline drift removal. A) R2 and mean blink duration (EOGv signal from subject #1) as a function of polynomial degree. The fifth order polynomial (R2 = 0.770) is marked with light blue and 20th order (R2 = 0.784) is marked with blue. The Figure shows that the R2 values and the mean blink durations are quite stationary for polynomials of order 5 to 25. B) Fitted fifth (light blue) and 20th order (blue) polynomials are plotted on the top of the raw signal (gray). The 20th order polynomial models the signal more closely than the 5th order polynomial. However, the difference between the fit (5th and 20th order) polynomials in the baseline denoised signals (C) and single blinks (D) is rather small. This figure shows that the baseline drift removal is fairly robust.

Back to article page