Automatic furniture arrangement. Make it home implementation.
Demo with bounding boxes + snip of csv database containing models.
Accessibility is checked through pathfinding over a voronoi subdivided version of the floor geometry. Each Iteration is scored and compared against current iteration.
First prototype of this case before swapping over to automatic optimization algorithm.
This is a 100% parametric bathroom, inspired by common online planners. Multiple variations are generated through mutation and iteration in TOP networks
These are then scored on clearance, accessbility and privacy. For privacy I cast rays from window panes to measure visibility. Each fixture has a "privacy tolerance". In this case, shower has frosted glass, so a higher privacy tolerance.
Scaling the bathroom model to bedrooms with procedural furniture, whose parameters can also be mutated.
This were the some of the proof of concepts I worked on under the Procedural 3D TETRA project @ DAE Research.
First is a simplified implenetation of MakeItHome: Automatic Optimization of Furniture Arrangement (Lap-fai yu et al.) in Houdini that runs inside of a dop network for iterative optimization of placement. I've also incorporated elements from other papers from P. Kan & h. Kaufman and T. Germer & M. Schwarz.
Models are loaded from a collection that is outlined in a csv file with various attributes and pairing settings that influence placing.
Second is a fully procedural bathroom generator and evaluator to minimize clearance overlaps, maximize privacy and accessibility. This generates a 3D maquette and 2D plan.
This was the older prototype, and when I tried scaling this up to other room types (bedrooms, kitchens), it was restrictive for rooms like living rooms. So I swapped over to automatic optimization for more complex room setups.
