Creating a supervision tool in Condition Based Monitoring (CBM) for manufacturing processes by means of a pattern recognition approach, with emphasis on the feature extraction and classification is usually a difficult task. In particular manufacturing methods like drilling, turning, milling, boring and grinding are of concern for the discussion. The issue of machine tool downtime and degraded productivity and production costs continues to plague the industry and thus urge for reliable CBM systems enabling to predict or detect vibration, estimating tool wear and detect tool breakage. Extracting relevant information and choosing a suitable classifier is far from trivial for a given CBM scenario and requires knowledge of the process involved. This paper will discuss some common techniques used and also aim to indicate possible new approaches utilizing emerging techniques from other disciplines. In particular, nonlinear techniques such as Local Binary Pattern (LBP) and variants thereof are investigated as possible techniques for feature extraction. Generative as well as discriminative classifiers are also discussed.