An Approach to Design Sketch Modelling
Author: Lim, Sungwoo
Supervisor: Duffy, Alex H. B.; Lee, B. S.
Institution: University of Strathclyde, Department of Design, Manufacture and Engineering Management, CAD Centre, Glasgow, Scotland, UK
Visualisation of conceptual models, in the form of sketches, is important for a more flexible and dynamic design. Many designs start with sketches, and it is of crucial importance to integrate the sketching activities into the CAD environment because of the advantages of sketching during conceptual design. Much work to date on sketch recognition and computeraided sketching (CAS) systems is proof of this. With the increase of an interest about the importance of a conceptual design stage, the interest about vagueness also started to increased by a few researchers [1-3]. However, current methods, especially in a sketch recognition approach, do not offer any method of representing and managing Vague Geometric Information (VGI) often found in conceptual ideas.
Therefore, an alternative novel approach Sketch Modelling (SM) based on a hierarchical structure and probabilistic method, and their associated issues have been established. The overall idea of SM is to support the designer to model a sketch with a representation and maintenance of the VGI of a sketch based conceptual design. SM consists of six main foundations: stroke analysis, 2D and 3D primitive shape probability calculation, hierarchical probability distribution, three-dimensional conversion, 2D and 3D Alternative Shape Type (AST) probability calculations, and hierarchical clustering with VGI. These foundations are formalised as computational algorithms to enable the SM via a computer based prototype system. A prototype system, I-MAGI, has been developed to evaluate the realisation of SM in conceptual design support. The functionality of I-MAGI has been illustrated throughout the evaluation with a worked example. Further, the strengths and weaknesses of SM have been discussed and, finally the future work to enhance the capability of SM and the functionality of I-MAGI in conceptual design support has been identified.