Research on accuracy and optimization for some baseline removal algorithms for high-throughput experiments

This article describes data processing for background removal, peak matching in spectrum analyses for experiments such as high-throughput experiments, which is subtracting background function from original data. We tested algorithms such as polynomial method, Whittaker-smoothing-based method, spline and morphological method, and make comparison among these common-used background removal algorithm. Using variable control for main parameters in each algorithm, and Euclidean norm for measuring the distance between original data and baseline function. We get the conclusion that morphological takes advantage in that its baseline function is nearest to original data. By analyzing theory, factor choosing and effectiveness, it is clear that regional graphic procedure and segment procedure are more effective. So further experience aim is determined.