- Title
- Towards the simulation of three dimensional image forms using a graphical genetic algorithm
- Creator
- Fairall, R. G.; Herbert, R.; Summons, P.
- Relation
- 18th World IMACS Congress and MODSIM09 International Congress on Modelling and Simulation. The 18th World IMACS Congress and MODSIM09 International Congress on Modelling and Simulation: Proceedings (Cairns, Qld 13-17 July, 2009) p. 4416-4422
- Relation
- http://www.mssanz.org.au/modsim09/authorsE-G.htm#f
- Publisher
- Modelling and Simulation Society of Australia and New Zealand and International Association for Mathematics and Computers in Simulation
- Resource Type
- conference paper
- Date
- 2009
- Description
- This paper examines some of the problems confronted in evolving a population of three dimensional computer generated shapes according to a set of desired constraints. The eventual aim is to automatically search a design or problem space for a viable solution or range of solutions to 3-D shapes. Such shapes are often used in generating computer images. Such images are needed in engineering and science and are useful in visualizing output from computer models. There may be applications within the domains of aesthetic design. Previously, most of the successful attempts to evolve three-dimensional computer generated form have used systems of genetic encoding with a high degree of complexity in the link between the Genotype (the genetic endowment of the individual solution) and the Phenotype (the constellation of all available individual solutions), and have mapped the complete volume of the shape. An approach described is to encode the shape surface. This has the potential to considerably simplify the mapping of the Phenotype to Genotype. The new approach is inspired by the principle behind the operation of a “shape grabber”. In this a mathematical array consisting of a “point cloud” of Cartesian coordinates is created corresponding to those mapped on the surface of a real object. The former can then be used by a CAD program to plot a polygonal mesh and produce a 3-D representation of the original object. The new approach uses a direction vector, the magnitude of which becomes a floating point “gene” within the Genetic algorithm. The endpoint of the latter defines a point within the point cloud. An enhancement of this model could use an evolved Lindenmayer System to define a point in the cloud. This study compares the Volumetric and Surface Models of shape encoding and surface evolution and points to new exciting possibilities.
- Subject
- three dimensional images; computer generated shapes; 3D forms; shape encoding
- Identifier
- http://hdl.handle.net/1959.13/919463
- Identifier
- uon:8876
- Identifier
- ISBN:9780975840078
- Language
- eng
- Reviewed
- Hits: 2493
- Visitors: 2696
- Downloads: 2
Thumbnail | File | Description | Size | Format |
---|