New haptic syringe device for virtual angiography training

Publication date: Available online 1 April 2019

Source: Computers & Graphics

Author(s): Dongjin Huang, Pengbin Tang, Xianglong Wang, Tao Ruan Wan, Wen Tang

Abstract

Angiography is an important minimally invasive diagnostic procedure in endovascular interventions. Effective training for the procedure is expensive, time consuming and resource demanding. Realistic simulation has become a viable solution to addressing such challenges. However, much of previous work has been focused on software issues. In this paper, we present a novel hardware system-an interactive syringe device with haptics as an add-on hardware component to 3D VR angiography training simulator. Connected to a realistic 3D computer simulation environment, the hardware component provides injection haptic feedback effects for medical training. First, we present the design of corresponding novel electronic units consisting of many design modules. Second, we describe a curve fitting method to estimate injection dosage and injection speed of the contrast media based on voltage variation between the potentiometer to increase the realism of the simulated training. A stepper motor control method is developed to imitate the coronary pressure for force feedback of syringe. Experimental results show that the validity and feasibility of the new haptic syringe device for achieving good diffusion effects of contrast media in the simulation system. A user study experiment with medical doctors to assess the efficacy and realism of proposed simulator shows good outcomes.

Full article: New haptic syringe device for virtual angiography training

Modeling fractures and cracks on tree branches

Publication date: Available online 1 April 2019

Source: Computers & Graphics

Author(s): Liuming Yang, Meng Yang, Gang Yang

Abstract

Wood is a complex biological material. Due to the anisotropy of wood, the difference in the structure and physical-mechanical properties of different tree species, and the irregularity and complexity of the internal structure of tree branches, it is challenging to model wood fractures caused by natural disasters and man-made damage. In this paper, we propose a user-controllable procedural modeling algorithm simulating the fractures based on the theory of wood science and wood fracture mechanics. It can also simulate cracks caused by wood shrinkage. The roughness of the corresponding area on the fracture surface is adjusted by setting the parameters of different material properties of pith, heartwood, sapwood, and bark; when the branches are broken, the fracture surface of the wood area under tension is treated as uneven and full of spines, and the pressure part is treated as a more regular section; the wood composed of different types of plant cells is represented by a group of particles. Various realistic cracks on wood fracture surface can be simulated by setting different strength and shrinkage rate in axial, radial and tangential direction. We illustrate our approach with various models from several tree species. Experimental results indicate that our approach can be used to model fractures and cracks on tree branches efficiently and realistically.

Full article: Modeling fractures and cracks on tree branches

PortraitNet: Real-time portrait segmentation network for mobile device

Publication date: Available online 4 April 2019

Source: Computers & Graphics

Author(s): Song-Hai Zhang, Xin Dong, Jia Li, Ruilong Li, Yong-Liang Yang

Abstract

Real-time portrait segmentation plays a significant role in many applications on mobile device, such as background replacement in video chat or teleconference. In this paper, we propose a real-time portrait segmentation model, called PortraitNet, that can run effectively and efficiently on mobile device. PortraitNet is based on a lightweight U-shape architecture with two auxiliary losses at the training stage, while no additional cost is required at the testing stage for portrait inference. The two auxiliary losses are boundary loss and consistency constraint loss. The former improves the accuracy of boundary pixels, and the latter enhances the robustness in complex lighting environment. We evaluate PortraitNet on portrait segmentation dataset EG1800 and Supervise-Portrait. Compared with the state-of-the-art methods, our approach achieves remarkable performance in terms of both accuracy and efficiency, especially for generating results with sharper boundaries and under severe illumination conditions. Meanwhile, PortraitNet is capable of processing 224 × 224 RGB images at 30 FPS on iPhone 7.

Full article: PortraitNet: Real-time portrait segmentation network for mobile device

Data-driven 3D human head reconstruction

Publication date: Available online 4 April 2019

Source: Computers & Graphics

Author(s): Huayun He, Guiqing Li, Zehao Ye, Aihua Mao, Chuhua Xian, Yongwei Nie

Abstract

This paper proposes a framework for reconstructing 3D human head models from a single image. Firstly, a preprocessing system for images is designed. The image is automatically segmented using convolutional neural network. The Gabor filter is used to extract the direction of the hair, and a optional manual interaction is used to modify the results. Secondly, the parameters of the FLAME face parametric model are solved by using landmark points as constraint. The high-frequency information of the face image is used to enhance the detail, and the texture map is complemented by the albedo parameterized model. Finally, a strip mesh hair database is constructed, and the hair in the image is reconstructed by using this database and the information extracted from the image.

Full article: Data-driven 3D human head reconstruction

Bidirectional Recurrent Autoencoder for 3D Skeleton Motion Data Refinement

Publication date: Available online 6 April 2019

Source: Computers & Graphics

Author(s): Shujie Li, Yang Zhou, Haisheng Zhu, Wenjun Xie, Yang Zhao, Xiaoping Liu

Abstract

In this paper, we propose a novel 3D skeleton human motion data refinement method that is based on a bidirectional recurrent autoencoder (BRA). The BRA has two main characteristics: (1) the motion manifold is extracted by a bidirectional long short-term memory recurrent neural network (B-LSTM-RNN) and (2) apart from statistical information of motion data, kinematic information including smoothness and bone length constrain, are also simultaneously exploited with noisy-clean motion pairs. Using a bidirectional LSTM unit, which is more suitable for time series and can infer information from the data in both time directions, our autoencoder extracts a manifold that can exploit the spatial and temporal relationships between previous and subsequent motion data. As a result, the refined data that are projected by the decoder from the motion manifold have much lower reproduction error. Furthermore, owing to the consideration of kinematic information, our reproduced motion data are of higher visual quality, while preserving positional precision. The proposed method is not action-specific and can handle a wide variety of noise types. The proposed method does not require the noise amplitude, which may be unknown in many scenarios, as a priori knowledge. Extensive experimental results demonstrate that our method outperforms several state-of-the-art methods.

Full article: Bidirectional Recurrent Autoencoder for 3D Skeleton Motion Data Refinement

GPU-based efficient computation of power diagram

Publication date: Available online 28 March 2019

Source: Computers & Graphics

Author(s): Zheng Liping, Gui Zhiqiang, Cai Ruiwen, Fei Yue, Zhang Gaofeng, Xu Benzhu

Abstract

Power diagram has been widely used in graphics and engineering. In the computation process for the centroidal capacity constrained power diagram, the geometrical construction of the power diagram is time consuming, and accounts for more than 50% of the total computing time, which severely affects the efficiency. A novel GPU-based power diagram construction algorithm is proposed to improve the efficiency. We introduce the jump flooding algorithm for parallelly rendering the power diagram, and present an approach for extracting the geometrical vertices and edges to generate the power diagram. Then, to accelerate the power diagram computation process, we propose a GPU–CPU hybrid algorithm by coupling the existing CPU-based algorithm with our proposed GPU algorithm. Furthermore, to accelerate further, a pure GPU algorithm is proposed. It utilizes the discrete Lloyd’s algorithm for centroidal constraints and a GPU-based analytical algorithm for weights and capacities. Experiment results exhibit that our proposed GPU algorithms can improve the efficiency of the power diagram construction by several orders of magnitude.

Full article: GPU-based efficient computation of power diagram

slice2mesh: A meshing tool for the simulation of additive manufacturing processes

Publication date: Available online 29 March 2019

Source: Computers & Graphics

Author(s): Marco Livesu, Daniela Cabiddu, Marco Attene

Abstract

Accurately simulating additive manufacturing (AM) processes is useful to predict printing failures and test 3D printing without wasting precious resources, both in terms of time and material. In AM the object to be fabricated is first cut into a set of slices aligned with the build direction, and then printed, depositing or solidifying material one layer on top of the other. To guarantee accurate simulations, it is therefore necessary to encode the temporal evolution of the shape to be printed within the simulation domain. We introduce slice2mesh, to the best of our knowledge the first software capable of turning a sliced object directly into a volumetric mesh. Our tool inputs a set of slices and produces a tetrahedral mesh that endows each slice in its connectivity. An accurate representation of the simulation domain at any time during the print can therefore be easily obtained by filtering out the slices yet to be processed. slice2mesh also features a flexible mesh generation system for external supports, and allows the user to trade accuracy for simplicity by producing approximate simulation domains obtained by filtering the object in slice space.

Graphical abstract

Graphical abstract for this article

Full article: slice2mesh: A meshing tool for the simulation of additive manufacturing processes

A comparison of methods for gradient field estimation on simplicial meshes

Publication date: Available online 30 March 2019

Source: Computers & Graphics

Author(s): Claudio Mancinelli, Marco Livesu, Enrico Puppo

Abstract

The estimation of the differential properties of a function sampled at the vertices of a discrete domain is at the basis of many applied sciences. In this paper, we focus on the computation of function gradients on triangle and tetrahedral meshes. We study one cell-based method (the standard the facto), plus three vertex-based methods. Comparisons regard accuracy, ability to perform on different domain discretizations, and efficiency. We performed extensive tests and provide an in-depth analysis of our results. Besides some common behaviour, we found that some methods perform better than others, considering both accuracy and efficiency. This directly translates to useful suggestions for the implementation of gradient estimators in research and industrial code.

Graphical abstract

Graphical abstract for this article

Full article: A comparison of methods for gradient field estimation on simplicial meshes

Mill and fold: Shape simplification for fabrication

Publication date: May 2019

Source: Computers & Graphics, Volume 80

Author(s): Alessandro Muntoni, Stefano Nuvoli, Andreas Scalas, Alessandro Tola, Luigi Malomo, Riccardo Scateni

Abstract

We introduce a pipeline for simplifying digital 3D shapes and fabricate them using 2D polygonal flat parts. Our method generates shapes that, once unfolded, can be fabricated with CNC milling machines using special tools called V-Grooves. These tools create V-shaped furrows at given angles depending on the shape of the used tool. Milling the edges of each flat facet simplifies the manual assembly, which consists only in folding adjacent facets at a constrained angle. Our method generates simplified shapes where every dihedral angle between adjacent facets belongs to a restricted set, thus making the assembly process quicker and more straightforward. Firstly, our method automatically computes a simplified version of the input model, using the marching cubes algorithm on the original mesh and iteratively performing local changes on the resulting triangle mesh. The user can then perform an additional manual simplification to remove unwanted facets. Finally, an unfolding algorithm, which takes into account the thickness of the material, flattens the polygonal facets onto the 2D plane, so that a CNC milling machine can fabricate it from a sheet of rigid material.

Graphical abstract

Graphical abstract for this article

Full article: Mill and fold: Shape simplification for fabrication

Interactive design of castable shapes using two-piece rigid molds

Publication date: Available online 15 March 2019

Source: Computers & Graphics

Author(s): Oded Stein, Alec Jacobson, Eitan Grinspun

Abstract

Two-piece rigid molds are particularly amenable to mass-manufacturing. Arbitrary objects cannot be realized with such a mold, because the cast will collide with the mold during removal. Most shapes are far away from being castable, i.e., realizable with such a two-piece rigid mold. When starting with an arbitrary shape, large deformations are needed to produce castable shapes. Such large deformations require user interaction together with a design process that is aware of fabrication constraints. We present such a design tool to generate two-piece rigid molds separated by a planar cut for a wide variety of shapes. Our casts can be produced in one single piece from a rigid, reusable mold and do not require further assembly after casting. We use a castability energy that can be optimized with gradient-based methods combined with an elastoplastic deformation based on the as-rigid-as-possible method. The tool enables a designer to deform an input shape significantly and arrive at an output shape that is castable.

Graphical abstract

Graphical abstract for this article

Full article: Interactive design of castable shapes using two-piece rigid molds