Special Section on SIBGRAPI 2014Sketch-based modeling and adaptive meshes
Graphical abstract
Introduction
Sketches are the most direct way to communicate shapes: humans are able to associate complex shapes with few curves. However, sketches do not have complete shape information, and the information sketches do provide is often inexact; thus, ambiguities are natural. On the other hand, to create, edit, and visualize shapes using computers, we need precise mathematical information, such as a function formula or a triangle mesh. The problem of how to model shapes using sketches can be formulated as how to fill the missing information about the model. In the last 15 years, sketch-based modeling (SBM) has become a well established research area, encompassing work in different domains, such as computer vision, human–computer interaction, and artificial intelligence [1]. However, this body of work lacks a more theoretical approach on how to build a sketch-based modeling system for a given application. In contrast, in this work we introduce a framework tailored for sketch-based surface modeling (SBSM) taking advantage of adaptive meshes. Based on the proposed framework we present and discuss two different sketch-based modeling systems.
We advocate that SBSM systems must be suited to each specific application: the specificities of a certain field require suitable mathematical representations for the domain model, and this plays a central role in the characterization of SBSM applications. However, there are common requirements in many SBSM applications that can be abstracted to guide the definition of specific representations for specific domains. These requirements have three main aspects: (1) dynamic – the surface will change during the modeling process; (2) interactive – the user must be able to see the model changing with interactive response and feedback; (3) controlled freedom – some applications have specific modeling rules and the systems must be able to incorporate these rules to guide the user in building a correct model, without losing flexibility.
Adaptive meshes are generally associated with the ability to produce detailed complex models using a smaller mesh. However, our proposed framework is based on adaptive meshes because they can be dynamic and enable rapid updates with local control. Different schemes of adaptive meshes can be used to create a system using our framework; indeed, the choice of the scheme must take into account the final application requirements, such as how to represent features, what changes of topology are allowed, and how smooth the models need to be. Fig. 1 shows an instance of a model built within our framework: a 4–8 adaptive mesh adapted to an implicit surface.
Based on the proposed framework we built two majorly different sketch-based modeling systems which are presented in this paper. The two sketch-based modeling systems that will be presented here are built using our proposed framework and have major differences. The first system is the Detail Aware Sketch-Based Surface Modeling (DASS, Section 5), which approaches a common problem in many SBSM systems: the lack of good control of global and local transformations. We created DASS to allow us to validate our proposed framework, exploring the limitations of a general system without a well defined task. To achieve the required control we developed a method to create atlas structures for adaptive meshes based on stellar operators [2]. The second system is the Geological Layer Modeler (Section 6), which is a sketch-based system specialized for geology that aims to help geophysicists to create subsurface models. This system is a good illustration of controlled freedom, where the sketch operators should be restricted to follow geological rules.
Section snippets
Related work
In the past decades there has been a large body of work in sketch-based surface modeling [3], [4], [5], [6], [7]. However, these systems are more concerned with the final results and do not consider the theoretical aspects of the mathematical surface representation used. We discuss below the main works on free-form sketch-based surface modeling that start from scratch under the light of its representations.
There are many ways to represent surfaces in . The most common and general are
Adaptive mesh overview
An adaptive mesh is a polygonal mesh that has the ability to create and remove vertices, edges, and faces following predefined rules. The creation process is called refinement and the deletion process is called simplification. An adaptive mesh scheme starts with a base mesh which is refined until it matches a stopcriterion. Usually this criterion is associated with a maximum threshold for some error metric. In summary, an adaptive mesh must have a base mesh, criteria for when to apply
Framework
The proposed framework enables system designers to build a sketch-based system that is interactive and has controlled freedom. Interactivity means that the system must be able to show how the model changes in interactive time. Controlled freedom means that some applications have specific modeling rules, and the system must be able to incorporate these rules to guide the modeler, but without losing flexibility. Moreover, the framework must be sufficiently general to be applied in different
Detail aware sketch-based surface modeling (DASS)
The main goal of DASS system prototype is to allow the user to control local modifications without changing parts of the model outside the region of interest, and keeping details coherent when large deformations are introduced. Hence, we advocate that decomposing the model representation into a base surface that supports different types of properties is a powerful tool for sketch-based surface modeling. Markedly, Blinn [30] introduces the idea of bump-mapping that stores geometric information
Geological layer modeler (GLaM)
We developed a sketch-based system for seismic interpretation and reservoir modeling (Fig. 14) based on the framework presented in Section 4. Most of the existing tools for seismic interpretation rely on the automatic extraction of horizons (interfaces between two rock layers) using segmentation algorithms. However, seismic data have a high level of uncertainty and noise which leads to mistakes in the horizon extraction. The main objective of the GLaM system is to enable the experts to directly
Conclusion and future work
We have presented two sketch-based systems to illustrate the flexibility of our framework. The adaptive mesh plays a central hole in this framework enabling rapid updates with local control. This work opens many interesting venues. One of the natural next steps is to use the framework in different domains and applications.
DASS system leaves many interesting open questions. This approach achieves good results, but it only allows us to work in a single plane. Since the base mesh is responsible
Acknowledgments
We would like to thank our colleagues for their useful discussions and advice, in particular to Nicole Sultanum. We also thank the anonymous reviewers for their careful and valuable comments and suggestions. This research was supported in part by the NSERC/Alberta Innovates Technology Futures (AITF)/Foundation CMG Industrial Research Chair Program in Scalable Reservoir Visualization and by grants from the Brazilian funding agencies CNPq and CAPES/PDEE.
References (44)
- et al.
Sketch-based modelinga survey
Comput Graph
(2009) - et al.
Sketch-based surface design using malleable curve networks
Comput Graph
(2012) Piecewise algebraic surface patches
Comput Aided Geom Des
(1985)- et al.
Freestylesculpting meshes with self-adaptive topology
Comput Graph
(2011) - et al.
Sculpting multi-dimensional nested structures
Comput Graph
(2013) - et al.
A simple and flexible framework to adapt dynamic meshes
Comput Graph
(2008) Using generic programming for designing a data structure for polyhedral surfaces
Comput Geom
(1999)- et al.
4-8 subdivision
Comput Aided Geom Des
(2001) - et al.
Adaptive multi-chart and multiresolution mesh representation
Comput Graph
(2014) A dynamic adaptive mesh library based on stellar operators
J Graph GPU Game Tools
(2004)
Fibermeshdesigning freeform surfaces with 3D curves
ACM Trans Graph
Structured annotations for 2D-to-3D modeling
ACM Trans Graph
Extending the CSG tree-warping, blending, and boolean operations in an implicit surface modeling system
Comput Graph Forum
Sketch-based 3D-shape creation for industrial styling design
IEEE Comput Graph Appl
Feature based terrain generation using diffusion equation
Comput Graph Forum
Free-form sketching with variational implicit surfaces
Comput Graph Forum
Cited by (3)
A 3D modeling methodology based on a concavity-aware geometric test to create 3D textured coarse models from concept art and orthographic projections
2018, Computers and Graphics (Pergamon)Citation Excerpt :They assume that silhouettes of sub-parts of a 2D drawing from two the front and side views can provide enough information for 3D modeling. The framework of Vital-Brazil et al. [23] is based on a Detail Aware Sketch-Based Surface Modeling (DASS) system that allows the user to create meshes from a sketch and refine it by controlling local modifications. This overview is by no means exhaustive, and we refer the reader to the cited related works and the surveys of Olsen et al. [24] and Cordier et al. [25] for a more complete picture.
Forward to the special section on SIBGRAPI 2014
2015, Computers and Graphics (Pergamon)Sketch2Relief: Generating Bas-relief from Sketches with Deep Generative Networks
2020, Proceedings - International Conference on Tools with Artificial Intelligence, ICTAI