Abstract:
|
Tasks in architectural and interior design range from defining the building floor plans and
ensuring desired functionality, to deciding furnishing styles and arrangement choices; all to
best fit certain pre-established purposes. The process of design, as a whole, has remained
hard to master for computer-based optimization in general and for computational
intelligence approaches in particular. Some attempts to tackle different subfields of this
problem in a machine learning fashion have emerged over the last few years, aiming to
offer partial automatization of human tasks, personalized support for specialists in the field
and professional guidance for amateurs. In this thesis, we first present an overview of
current advances of computational intelligence in architectural science with a focus on
interior design. We describe various learning models applied to interior design challenges
such as furniture type selection, style compatibility, furniture arrangement, or ornamental
decoration. The core of the thesis is devoted to report ongoing research towards the
development of a commercial, robust and scalable solution for automatic furniture
arrangement, given a room plan. We propose two probabilistic models to be used in the
complex problem of furnishing bedrooms. The first resides in a Bayesian Network based
approach for the automatic generation of the number and types of furniture entities to
occupy the new space, namely the occurrence model. The second one, called
arrangement model, deals with learning different commonly met sets of items
interconnected within the same space and estimating their relative positions with GMMs.
Both models heavily contribute to the main goal of achieving a 3D planner for bedrooms,
but their genericity allows other types of interiors to be modeled through the same process. |