Research Article

Automation of the Peak Fitting Method in Bone FTIR Microspectroscopy Spectrum Analysis: Human and Mice Bone Study

Figure 3

Visual representation of the algorithm. Evolution of the negative part of the second derivative of the bone spectrum between the MMA peak of interest boundaries (1706 cm−1 and 1826 cm−1) through the iterations. Each iteration corresponds to the subtraction of the MMA-embedded bone spectrum minus the weighted MMA spectrum (the weight changes with each iteration). Note. An iteration is not always better than the previous one (e.g., iterations 1 and 2), but it tends nonetheless to converge to the expected result (e.g., iteration 22), which is the convexity of the spectrum between the 1706 cm−1 and 1826 cm−1 boundaries. The final result of the MMA removal can be observed in Figure 1 (the MMA contribution, including the 1730 cm−1 peak, in the bone spectrum is removed).