Features are information clusters or “chunks”. According to cognitive psychologists information chunks are central to human cognition; they are structures for encoding and storing knowledge in memory. Features are used in pattern recognition, human communications, reasoning and other cognitive tasks. In the context of engineering artifacts, form features are chunks of related facts and shape characteristics of an engineering artifact.

The major tasks in feature recognition from CAD models are:

• Identify & characterize features of interest for a given application
• Formalize features in terms of CAD model entities and relations
• Create intermediate data structures from CAD model to facilitate recognition
• Search the intermediate model to match the key feature property
• Test additional properties from adjacent entities in geometric model
• Extract recognized features from the geometric model
• Extract feature parameters (e.g., hole diameter, pocket depth)
• Combine simple features to get higher level features

ASU DAL is well known for its pioneering research in parametric and feature based modeling, both design by features and feature recognition. The evolution of feature recognizers in DAL is listed below:

• Rule based early feature recognizers – 1988
• Volume decomp – recomp recognizer – 1994
• Machining algebra for mapping volumes to processes - 1995
• Feature independent recognition algorithm – 1996
• Interactive definition of form features – 1996
• 3-axis milling graph based method – 1997
• Integration of Design by feature & Feature Recognition - 1998
• Universal hint based method – 2002
• Feature recognizer inside commercial CAD system (I_DEAS) – 2002
• Free form feature recognizer – 2003
• Prismatic UDF Recognizer – 2004
• Feature Tutor© for N-Rep© UDF library creation – 2005
• Turning features UDF recognizer from Mill-Turn parts – 2006
• Assembly features based on screw theory - (in progress)