SUPERVISED LEARNING // ROOF FORM RECOGNITION

Aims:

The project aims to train the neural network to distinguish between four different roof forms, including, shed roof, hip roof, barrel vault and flat roof.

Learning Objectives:

  • Understand a range of machine learning methods
  • Be able to couple parametric models with machine learning tools
  • Setup supervised learning using Grasshopper
  • Be aware of data preprocessing necessary for successful training

Strategies:

  • The number of faces
  • The sum of XY components of the normal vector of the roof surface(s)
  • The Z component of the unitized normal vector of the roof surface(s)
  • The volume difference between the mass and the bounding box of the mass
  • The change in tangent in the sectional profile

 

 

 

 

 

 

 

 

 

 

Roof Form Recognition is a project of IAAC, Institute for Advanced Architecture of Catalonia developed in the Master in Advanced Architecture 2019/20 by Timothy Ka Kui Lam and Faculty: Mateusz Zwierzycki