String-Based Synthesis of Structured Shapes

Javor Kalojanov1         Isaak Lim1         Niloy Mitra2         Leif Kobbelt1

1Visual Computing Institute, RWTH Aachen University         2University College London        



We propose a novel method to synthesize geometric models from a given class of context-aware structured shapes such as buildings and other man-made objects. The central idea is to leverage powerful machine learning methods from the area of natural language processing for this task. To this end, we propose a technique that maps shapes to strings and vice versa, through an intermediate shape graph representation. We then convert procedurally generated shape repositories into text databases that, in turn, can be used to train a variational autoencoder. The autoencoder enables higher level shape manipulation and synthesis like, for example, interpolation and sampling via its continuous latent space. We provide project code and pre-trained models.

    journal = {Computer Graphics Forum},
    title = {{String-Based Synthesis of Structured Shapes}},
    author = {Javor Kalojanov and Isaak Lim and Niloy Mitra and Leif Kobbelt},
    pages = {027-036},
    volume= {38},
    number= {2},
    year = {2019},
    note = {\URL{}},
    DOI = {10.1111/cgf.13616},