|논문명|A Survey of 2D and 3D Shape Descriptors | |-|-| |저자(소속)|Ismail Khalid Kazmi, (Bournemouth University )| |학회/년도| IEEE 2013, 논문| |키워드| | |참고|| |코드||

정의 : a shape descriptor is a simplified representation of a 2D or 3D shape in the form of a vector containing a set of numerical values or a graphlike structure used to describe the shape geometrically or topologically

The following common characteristics of an effective shape descriptor have been proposed in [1], [2], [3], [4].

• a) Discriminative accuracy: To accurately distinguish one shape from another based on subtle differences
• b) Transformation (translation, scaling, and rotation) invariance: Also known as pose normalization
• c) Robustness against model degeneracies /roughness
• d) Uniqueness: Each shape descriptor must be uniquely coupled with a unique shape
• e) Performance and memory efficient
• f) Partial matching: robust against incomplete shapes
• g) Insensitive to noise: Small changes in the shape to lead to small changes in the shape descriptor