Sentence segmentation is typically a first step in text normalization. Modern systems use a combination of complex regular expressions and machine learning classifiers for sentence segmentation.