Curve and surface fitting by implicit polynomials: Optimum degree finding and heuristic refinement

Publication date: October 2017
Source:Computers & Graphics, Volume 67
Author(s): Ruben Interian, Juan M. Otero, Celso C. Ribeiro, Anselmo A. Montenegro
Finding an implicit polynomial that fits a set of observations X is the goal of many researches in recent years. However, most existing algorithms assume the knowledge of the degree of the implicit polynomial that best represents the points. This paper presents two main contributions. First, a new distance measure between X and the implicit polynomial is defined. Second, this distance is used to define an algorithm able to find the degree of the polynomial needed for the representation of the data set. The proposed algorithm is based on the idea of gradually increase the degree, while there is an improvement in the smoothness of the solutions. The experiments confirm the validity of the approach for the selected 2D and 3D datasets.

Full article: Curve and surface fitting by implicit polynomials: Optimum degree finding and heuristic refinement

Combining traditional and indirect augmented reality for indoor crowded environments. A case study on the Casa Batlló museum

Publication date: Available online 28 September 2017
Source:Computers & Graphics
Author(s): Jesus Gimeno, Cristina Portalës, Inmaculada Coma, Marcos Fernández, Bibiana Martínez
Mobile audio-guides are currently being used in museums to enhance the visitors’ experience. Over the past few years, some Augmented Reality (AR) solutions have been explored in this context, making it possible to provide augmented visual and sonic stimuli through the use of smartphones. However, the special conditions of museums (e.g. artificial markers cannot be used, small, crowded rooms, etc.) and the limited computational capacity of smartphones mean that there are important restrictions in the use of this technology. In this paper we introduce a novel mobile augmented guide for the Casa Batlló museum (Barcelona, Spain) which is based on a combination of traditional and indirect AR. Through the proposed solution, we further explore the use of indirect AR for the interior of buildings as, at present, this technology is usually used for outdoor environments. The developed application is currently being used by Casa Batlló visitors. Retrieval of user experience shows some of the benefits of the proposed solution and opens the door for other similar solutions in museums.

Full article: Combining traditional and indirect augmented reality for indoor crowded environments. A case study on the Casa Batlló museum

Better initialization for regression-based face alignment

Publication date: Available online 8 August 2017
Source:Computers & Graphics
Author(s): Hengliang Zhu, Bin Sheng, Zhiwen Shao, Yangyang Hao, Xiaonan Hou, Lizhuang Ma
Regression-based face alignment algorithms predict facial landmarks by iteratively updating an initial shape, and hence are always limited by the initialization. Usually, the initial shape is obtained from the average face or by randomly picking a face from the training set. In this study, we discuss how to improve initialization by studying a neighborhood representation prior, leveraging neighboring faces to obtain a high-quality initial shape. In order to further improve the estimation precision of each facial landmark, we propose a face-like landmark adjustment algorithm to refine the face shape. Extensive experiments demonstrate our algorithm achieves favorable results compared to the state-of-the-art algorithms. Moreover, our algorithm achieves a smaller normalized mean error than the human performance (5.54% vs. 5.6%) on the challenging dataset the Caltech Occluded Faces in the Wild (COFW).

Full article: Better initialization for regression-based face alignment

On the visibility locations for continuous curves

Publication date: August 2017
Source:Computers & Graphics, Volume 66
Author(s): Sarang Joshi, Yoshida Rao, Bharath Ram Sundar, Ramanathan Muthuganapathy
The problem of determining visibility locations (VLs) on/inside a domain bounded by a planar C 1-continuous curve (without vertices), such that entire domain is covered, is discussed in this paper. The curved boundary has been used without being approximated into lines or polygons. Initially, a few observations regarding the VLs for a curved boundary have been made. It is proposed that the set of VLs required to cover the domain be placed in a manner that the VLs and the lines connecting them form a spanning tree. Along with other observations, an algorithm has been provided which gives a near optimal number of VLs. The obtained number of VLs is then compared with a visibility disjoint set, called as witness points, to obtain a measure of the ‘nearness’ of the number of VLs to the optimum. The experiments on different curved shapes illustrate that the algorithm captures the optimal solution for many shapes and near-optimal for most others.

Full article: On the visibility locations for continuous curves

Generalized drag force for particle-based simulations

Publication date: Available online 21 September 2017
Source:Computers & Graphics
Author(s): Christoph Gissler, Stefan Band, Andreas Peer, Markus Ihmsen, Matthias Teschner
Computing the forces acting from a surrounding air phase onto a particle-based fluid or rigid object is challenging. Simulating the air phase and modeling the interactions using a multiphase approach is computationally expensive. Furthermore, stability issues may arise in such multiphase simulations. In contrast, the effects from the air can be approximated efficiently by employing a drag equation. Here, for plausible effects, the parameterization is important but challenging. We present a drag force discretization based on the drag equation that acts on each particle separately. It is used to compute the effects of air onto particle-based fluids and rigid objects. Our presented approach calculates the exposed surface area and drag coefficient of each particle. For fluid particles, we approximate their deformation to improve the drag coefficient estimation. The resulting effects are validated by comparing them to the results of multiphase SPH simulations. We further show the practicality of our approach by combining it with different types of SPH fluid solvers and by simulating multiple, complex scenes.

Full article: Generalized drag force for particle-based simulations

Enhanced vector field visualization via Lagrangian accumulation

Publication date: Available online 2 August 2017
Source:Computers & Graphics
Author(s): Lei Zhang, Duong Nguyen, David Thompson, Robert Laramee, Guoning Chen
In this paper, we revisit the Lagrangian accumulation process that aggregates the local attribute information along integral curves for vector field visualization. Similar to the previous work, we adopt the notation of the Lagrangian accumulation field or A field for the representation of the accumulation results. In contrast to the previous work, we provide a more in-depth discussion on the properties of A fields and the meaning of the patterns exhibiting in A fields. In particular, we revisit the discontinuity in the A fields and provide a thorough explanation of its relation to the flow structure and the additional information about the flow that it may reveal. In addition, other remaining questions about the A field, such as its sensitivity to the selection of integration time, are also addressed. Based on these new insights, we demonstrate a number of enhanced flow visualizations aided by the accumulation framework and the A fields, including a new A field guided ribbon placement, an A field guided stream surface seeding and the visualization of particle-based flow data. To further demonstrate the generality of the accumulation framework, we extend it to the non-integral geometric curves (i.e., streak lines), which enable us to reveal information about the flow behavior other than those revealed by the integral curves. Finally, we introduce the Eulerian accumulation, which can reveal different flow behaviors than those revealed by Lagrangian accumulation. In summary, we believe that Lagrangian accumulation and the resulting A fields offer a valuable means to investigate flow behavior complementary to the current state-of-the-art techniques.

Full article: Enhanced vector field visualization via Lagrangian accumulation

Height-field construction using cross contours

Publication date: August 2017
Source:Computers & Graphics, Volume 66
Author(s): Tuan Minh Bui, Junho Kim, Yunjin Lee
We developed an interactive modeling tool for constructing a height field from the cross contours in a line drawing, added to provide an illusion of depth. Our approach places no specific constraints on the artist and can handle freely drawn shapes with partially structured or partially connected cross contours. In our system, the user first draws cross contours across the interior boundaries of the object. The system then extracts the intersection points of the contours and derives 2D polygons by connecting consecutive intersection points with straight lines. These 2D polygons are treated as the projection of a polygonal surface approximating the shape of the object. The 3D edge vectors and 3D normals of the approximated polygon surface are then estimated while considering the perceptions of the viewer and the goals of the designer. Based on the relative height values obtained from the 3D edge vectors and the co-planarities of the curve segments, a height field is generated in the image space that reasonably matches the cross contours of the input drawing. Experimental results demonstrated the ability of the system to generate height fields from a wide range of cross contours. We also showed that our system can be used to apply lighting effects to sketches and allows local shape deformation based on the constructed height field.

Full article: Height-field construction using cross contours

Exclusive grouped spatial hashing

Publication date: Available online 25 August 2017
Source:Computers & Graphics
Author(s): Weiwei Duan, Jianxin Luo, Guiqiang Ni, Bin Tang, Qi Hu, Yi Gao
A novel multidimensional hashing scheme, named the Exclusive Grouped Spatial Hashing (EGSH), which compresses repetitive spatial data into several compact tables while retaining efficient random access, is presented. EGSH represents a multi-level hashing without any losses. Moreover, EGSH compresses a group of repetitive elements into the same entry of the hash tables, while it uses a coverage table to mark the corresponding hash tables of the compressed data. Although, prior hashing work is related to hash collisions mitigation, here a full use of these collisions is obtained and therefore the spatial data compression rate is improved. The performance of exclusive grouped spatial hashing is presented in 2D and 3D graphic examples.

Full article: Exclusive grouped spatial hashing

Retrieving indoor objects: 2D-3D alignment using single image and interactive ROI-based refinement

Publication date: Available online 24 August 2017
Source:Computers & Graphics
Author(s): Fuchang Liu, Shuangjian Wang, Dandan Ding, Qingshu Yuan, Zhengwei Yao, Zhigeng Pan, Haisheng Li
Given a single indoor image, this paper proposes an automatic retrieval system to estimate the best-matching 3D models with consistent style and pose. To support this system, we combine a deep CNN based object detection approach with a deformable part based alignment model. The key idea is to cast a 2D-3D alignment problem as a part-based cross-domain matching. We also provide an interactive refinement interface that allows users to browse models based on similarities and differences between shapes in user-specified regions of interest (ROIs). We demonstrate the ability of our system on numerous examples.

Full article: Retrieving indoor objects: 2D-3D alignment using single image and interactive ROI-based refinement

Embedding shapes with Green’s functions for global shape matching

Publication date: November 2017
Source:Computers & Graphics, Volume 68
Author(s): Oliver Burghard, Alexander Dieckmann, Reinhard Klein
We present a novel approach for the calculation of dense correspondences between non-isometric shapes. Our work builds on the well known functional map framework and investigates a novel embedding for the alignment of shapes. We therefore identify points with their Green’s functions of the Laplace–Beltrami operator, and hence, embed shapes into their own function space. In our embedding the L 2 distances are known as the biharmonic distances, so that our embedding preserves the intrinsic distances on the shape. In the novel embedding each point-to-point map between two shapes becomes and can be represented as an affine map. Functional constraints and novel conformal constraints can be used to guide the matching process.

Full article: Embedding shapes with Green’s functions for global shape matching