Inverse procedural modeling (IPM)

 

 

What is Inverse procedural modeling?

Proceduralization of provided 3D content. Extracting rules and/or parameters to create a generative system from given buildings, plants, clouds, cities, and worlds.

A Proceduralized World:

 

 

 

IPM Classification

It converts existing geometry into procedural models (see (a)) and can be reduced to two:

 – 1. inferring procedural model and its parameters, as shown in (b)

• Only low-level priors can be assumed

• Usually operates directly on the geometry

– 2. inferring procedural model parameters (i.e., rules are known), as shown in (c)

• High-level priors can be assumed

• PM system is given but its usage must be determined/searched

                                                   

                                                                                         (a)

  

                                                       (b)                                                                                               (c)

 

 

 

What are the current approaches?

• Having low-level priors:

        – Formulations to partition geometry into components and then extract patterns and rules.

        e.g. ref [4] and [2]

• Having high-level priors:

        – Guiding and controlling the content generation by discovering the optimal parameters and rules for a given target.

        e.g. ref [3] Guided proceduralization

 

 

Low-level priors：Inferring Shape Grammar

Inspiration: Compositional Modeling

Definition: – Geometric formulations with minimal assumptions

Properties of approaches:

        – Exploiting redundancy in shapes

        – Extracting building blocks & assembly rules

        – Assuming no semantic / domain knowledge

Two main steps:

       • Extracting Building Blocks

       • Finding Rules & Synthesis Algorithms

Challenges:

• Modeling: How to define building blocks?

    –Redundancy relates to symmetry

    –Define notion of redundancy

• Computation: Two aspects

    –Matching: Detect similar pieces

    –Segmentation: Define pieces

 

 

References:

[1] Demir I, Aliaga D G, Benes B. Proceduralization of Urban models[C]//2017 25th Signal Processing and Communications Applications Conference (SIU). IEEE, 2017: 1-4.

[2] Demir I, Aliaga D G, Benes B. Proceduralization for editing 3d architectural models[C]//2016 Fourth International Conference on 3D Vision (3DV). IEEE, 2016: 194-202.

[3] Demir I, Aliaga D G. Guided proceduralization: Optimizing geometry processing and grammar extraction for architectural models[J]. Computers & Graphics, 2018, 74: 257-267.

[4] Demir I, Aliaga D G, Benes B. Coupled segmentation and similarity detection for architectural models[J]. ACM Transactions on Graphics (TOG), 2015, 34(4): 1-11.

[5] Inverse procedural modeling of 3D models for virtual worlds | ACM SIGGRAPH 2016 Courses https://dl.acm.org/doi/10.1145/2897826.2927323