Freeform features are parameterizable shape parts that are used in the design of industrial products. The parametric nature of the feature allows a designer to quickly manipulate shape without having to precisely configure the geometry of the shape. However, in many cases, designers want to use existing shapes in their designs, for which features have not been defined. In this case part of the existing shape must be (re)interpreted as feature data. In this thesis, new theory and methodology is given that addresses the problem of freeform feature recognition, in which geometric data is interpreted as feature data, thereby opening up the possibility for the designer to modify the shape parametrically. The developed method uses an evolutionary procedure, in which a feature shape that is defined beforehand is matched to the geometric data a designer wants to interpret. The procedure mimics the process of natural evolution and âevolvesâ towards an optimal solution. Once a solution has been found, the structure of the predefined feature shape matches that of the geometric data and can be used to parametrically control the shape data. The developed freeform feature recognition method significantly increases the accuracy and efficiency with which existing shapes can be incorporated in a design, and increases the level of control the designer has over the shape.