A STUDY OF A SEARCH METHOD OF FACADE GLASS ATTRIBUTES BY AN AESTHETIC EVALUATION OF CG IMAGES APPLYING AN INTERACTIVE EVOLUTIONARY COMPUTATION
Daisuke Matsushita, Junzo Munemoto · Journal of Architecture and Planning (Transactions of AIJ) · 2004
We tried the method to search the glass attributes, which are color, reflectance and transmittance, of objective glass facade images by an aesthetic evaluation of ray-tracing CG applying an interactive genetic algorithm. To verify the method, we firstly used the existent CG image as an objective image and searched it. Resultant acquired image closely resembled the objective image and both glass attributes are also close for the most part. We applied the method to a hotel facade, and reached an image that satisfied an chooser's expectation highly. The availability of this method to search the glass at-tributes by combining human aesthetic evaluation, high reproducibility of CG and multiple search procedure of GA was shown.