MVCNN++: Computer-Aided Design Model Shape Classification and Retrieval Using Multi-View Convolutional Neural Networks

2020 ◽  
Vol 21 (1) ◽  
Author(s):  
Atin Angrish ◽  
Akshay Bharadwaj ◽  
Binil Starly
Keyword(s):  
Neural Networks ◽  
Computer Aided Design ◽  
Training Data ◽  
Shape Classification ◽  
Training Time ◽  
3D Cad ◽  
Computer Aided ◽  
Cad Models ◽  
Engineering Models ◽  
Aided Design

Abstract Deep neural networks (DNNs) have been successful in classification and retrieval tasks of images and text, as well as in the graphics domain. However, these DNNs algorithms do not translate to 3D engineering models used in the product design and manufacturing. This paper studies the use of multi-view convolutional neural network (MVCNN) algorithm enhanced by the addition of engineering metadata, for classification and retrieval of 3D computer-aided design (CAD) models. The proposed algorithm (MVCNN++) builds on the MVCNN algorithm with the addition of part dimension data, improving its efficacy for manufacturing part classification and yielding an improvement in classification accuracy of 5.8% over the original version. Unlike datasets used for 3D shape classification and retrieval in the computer graphics domain, engineering level description of 3D CAD models do not yield themselves to neat, distinct classes. Techniques such as relaxed-classification and prime angled cameras for capturing feature detail were used to address training data capture issues specific to 3D CAD models, along with the use of transfer learning to reduce training time. Our study has shown that DNNs can be used to search and discover relevant 3D engineering models in large public repositories, making 3D models accessible to the community.

scholarly journals Automatized Evaluation of Students’ CAD Models

2021 ◽  
Vol 11 (4) ◽  
pp. 145
Author(s):  
Nenad Bojcetic ◽  
Filip Valjak ◽  
Dragan Zezelj ◽  
Tomislav Martinec
Keyword(s):  
Computer Aided Design ◽  
Three Dimensional ◽  
Computer Application ◽  
Test Cases ◽  
Cad Model ◽  
Computer Solution ◽  
3D Cad ◽  
Computer Aided ◽  
Cad Models ◽  
Aided Design

The article describes an attempt to address the automatized evaluation of student three-dimensional (3D) computer-aided design (CAD) models. The driving idea was conceptualized under the restraints of the COVID pandemic, driven by the problem of evaluating a large number of student 3D CAD models. The described computer solution can be implemented using any CAD computer application that supports customization. Test cases showed that the proposed solution was valid and could be used to evaluate many students’ 3D CAD models. The computer solution can also be used to help students to better understand how to create a 3D CAD model, thereby complying with the requirements of particular teachers.

scholarly journals DEMOCRATISING DESIGN THROUGH SURROGATE MODEL CONVOLUTIONAL NEURAL NETWORKS OF COMPUTER AIDED DESIGN REPOSITORIES

2020 ◽  
Vol 1 ◽  
pp. 1285-1294
Author(s):  
J. Gopsill ◽  
S. Jennings
Keyword(s):  
Neural Networks ◽  
3D Printing ◽  
Convolutional Neural Networks ◽  
Real World ◽  
Surrogate Model ◽  
Computer Aided Design ◽  
Computer Aided ◽  
Cad Models ◽  
Knowledge Intensive ◽  
Aided Design

AbstractThe capability to manufacture at home is continually increasing with technologies, such as 3D printing. However, the ability to design products suitable for manufacture and use remains a highly-skilled and knowledge intensive activity. This has led to ‘content creators’ providing vast repositories of manufacturable products for society, however challenges remain in the search & retrieval of models. This paper presents the surrogate model convolutional neural networks approach to search and retrieve CAD models by mapping them directly to their real-world photographed counterparts.

A Methodology for Building Up an Infrastructure of Haptically Enhanced Computer-Aided Design Systems

2008 ◽  
Vol 8 (4) ◽  
Author(s):  
Weihang Zhu
Keyword(s):  
Computer Aided Design ◽  
Force Feedback ◽  
Haptic Rendering ◽  
Haptic Interface ◽  
Cad Systems ◽  
3D Cad ◽  
Cad System ◽  
Computer Aided ◽  
Cad Models ◽  
Aided Design

This paper presents an infrastructure that integrates a haptic interface into a mainstream computer-aided design (CAD) system. A haptic interface, by providing force feedback in human-computer interaction, can improve the working efficiency of CAD/computer-aided manufacturing (CAM) systems in a unique way. The full potential of the haptic technology is best realized when it is integrated effectively into the product development environment and process. For large manufacturing companies this means integration into a commercial CAD system (Stewart, et al., 1997, “Direct Integration of Haptic User Interface in CAD Systems,” ASME Dyn. Syst. Control Div., 61, pp. 93–99). Mainstream CAD systems typically use constructive solid geometry (CSG) and boundary representation (B-Rep) format as their native format, while internally they automatically maintain triangulated meshes for graphics display and for numerical evaluation tasks such as surface-surface intersection. In this paper, we propose to render a point-based haptic force feedback by leveraging built-in functions of the CAD systems. The burden of collision detection and haptic rendering computation is alleviated by using bounding spheres and an OpenGL feedback buffer. The major contribution of this paper is that we developed a sound structure and methodology for haptic interaction with native CAD models inside mainstream CAD systems. We did so by analyzing CAD application models and by examining haptic rendering algorithms. The technique enables the user to directly touch and manipulate native 3D CAD models in mainstream CAD systems with force/touch feedback. It lays the foundation for future tasks such as direct CAD model modification, dynamic simulation, and virtual assembly with the aid of a haptic interface. Hence, by integrating a haptic interface directly with mainstream CAD systems, the powerful built-in functions of CAD systems can be leveraged and enhanced to realize more agile 3D CAD design and evaluation.

Graph-Based Simplification of Feature-Based Three-Dimensional Computer-Aided Design Models for Preserving Connectivity

2015 ◽  
Vol 15 (3) ◽  
Author(s):  
Soonjo Kwon ◽  
Byung Chul Kim ◽  
Duhwan Mun ◽  
Soonhung Han
Keyword(s):  
Computer Aided Design ◽  
Three Dimensional ◽  
Evaluation Metrics ◽  
Prototype System ◽  
Cad Model ◽  
3D Cad ◽  
Computer Aided ◽  
Cad Models ◽  
Feature Based ◽  
Aided Design

The required level of detail (LOD) of a three-dimensional computer-aided design (3D CAD) model differs according to its purpose. It is therefore important that users are able to simplify a highly complex 3D CAD model and create a low-complexity one. The simplification of a 3D CAD model requires the application of a simplification operation and evaluation metrics for the geometric elements of the 3D CAD model. The evaluation metrics are used to select those elements that should be removed. The simplification operation removes selected elements in order to simplify the 3D CAD model. In this paper, we propose the graph-based simplification of feature-based 3D CAD models using a method that preserves connectivity. First, new evaluation metrics that consider the discrimination priority among several simplification criteria are proposed. Second, a graph-based refined simplification operation that prevents the separation of a feature-based 3D CAD model into multiple volumes is proposed. Finally, we verify the proposed method by implementing a prototype system and performing simplification experiments using feature-based 3D CAD models.

A New Family of Bricard-Derived Deployable Mechanisms

2016 ◽  
Vol 8 (3) ◽  
Author(s):  
Hailin Huang ◽  
Bing Li ◽  
Jianyang Zhu ◽  
Xiaozhi Qi
Keyword(s):  
Large Volume ◽  
Computer Aided Design ◽  
Degree Of Freedom ◽  
Single Degree Of Freedom ◽  
Single Degree ◽  
New Family ◽  
Revolute Joints ◽  
Computer Aided ◽  
Cad Models ◽  
Aided Design

This paper proposes a new family of single degree of freedom (DOF) deployable mechanisms derived from the threefold-symmetric deployable Bricard mechanism. The mobility and geometry of original threefold-symmetric deployable Bricard mechanism is first described, from the mobility characterstic of this mechanism, we show that three alternate revolute joints can be replaced by a class of single DOF deployable mechanisms without changing the single mobility characteristic of the resultant mechanisms, therefore leading to a new family of Bricard-derived deployable mechanisms. The computer-aided design (CAD) models are used to demonstrate these derived novel mechanisms. All these mechanisms can be used as the basic modules for constructing large volume deployable mechanisms.

Intelligent Automated Welding for Shipyard Applications

1992 ◽  
Vol 8 (02) ◽  
pp. 77-88
Author(s):  
S. Madden ◽  
H. H. Vanderveldt ◽  
J. Jones
Keyword(s):  
Neural Networks ◽  
Computer Technology ◽  
Computer Aided Design ◽  
Data Systems ◽  
Fully Integrated ◽  
Computer Aided Process Planning ◽  
Computer Aided ◽  
Process Operation ◽  
Aided Design ◽  
Computerized System

Computer Aided Process Planning (CAPP) integrated with Computer Aided Design (CAD) and Computer Aided Manufacturing (CAM) will form the basis of engineering/planning systems of the future. These systems will have the capability to operate in a paperless environment and provide highly optimized process operation plans. The WELDEXCELL System is a prototype of such a system for welding in shipyards. The paper discusses three significant computer technology advances which have been in into the WELDEXCELL prototype. First is a computerized system for allowing multiple knowledge sources (expert systems, humans, data systems, etc.) to work together to solve a common problem (the weld plan). This system is called a "blackboard." The second is a methodology for the blackboard to communicate to the human user. This interface includes full interactive graphics fully integrated to CAD as well as data searches and automatic completion of routine engineering tasks. The third is artificial neural networks (ANS's), which are based on biological neural networks (such as the human brain) and which can do neural reasoning tasks about difficult problems. ANS's offer the opportunity to model highly complex multivariable and nonlinear processes (for example, welding) and provide a means for an engineer to quantitatively assess the process and its operation.

Orthogonal Distance Fields Representation for Machine-Learning Based Manufacturability Analysis

2020 ◽  
Author(s):  
Aditya Balu ◽  
Sambit Ghadai ◽  
Soumik Sarkar ◽  
Adarsh Krishnamurthy
Keyword(s):  
Machine Learning ◽  
Computer Aided Design ◽  
Boundary Representation ◽  
Engineering Product ◽  
Distance Fields ◽  
Computer Aided ◽  
Cad Models ◽  
Product Design Process ◽  
3D Cnn ◽  
Aided Design

Abstract Computer-aided Design for Manufacturing (DFM) systems play an essential role in reducing the time taken for product development by providing manufacturability feedback to the designer before the manufacturing phase. Traditionally, DFM rules are hand-crafted and used to accelerate the engineering product design process by integrating manufacturability analysis during design. Recently, the feasibility of using a machine learning-based DFM tool in intelligently applying the DFM rules have been studied. These tools use a voxelized representation of the design and then use a 3D-Convolutional Neural Network (3D-CNN), to provide manufacturability feedback. Although these frameworks work effectively, there are some limitations to the voxelized representation of the design. In this paper, we introduce a new representation of the computer-aided design (CAD) model using orthogonal distance fields (ODF). We provide a GPU-accelerated algorithm to convert standard boundary representation (B-rep) CAD models into ODF representation. Using the ODF representation, we build a machine learning framework, similar to earlier approaches, to create a machine learning-based DFM system to provide manufacturability feedback. As proof of concept, we apply this framework to assess the manufacturability of drilled holes. The framework has an accuracy of more than 84% correctly classifying the manufacturable and non-manufacturable models using the new representation.

Paired Computer-Aided Design: The Effect of Collaboration Mode on Differences in Model Quality

2020 ◽  
Author(s):  
Hamza Arshad ◽  
Vrushank Phadnis ◽  
Alison Olechowski
Keyword(s):  
Computer Aided Design ◽  
Shared Control ◽  
Control Mode ◽  
Pair Programming ◽  
Design Teams ◽  
Model Quality ◽  
Computer Aided ◽  
Parallel Cad ◽  
Cad Models ◽  
Aided Design

Abstract We present the results of an experiment investigating two different modes of collaboration on a series of computer-aided design (CAD) tasks. Inspired by the pair programming literature, we anticipate that partners working in a fully synchronous collaborative CAD environment will achieve different levels of quality in CAD models depending on their mode of collaboration — one in which the pair is free to work in parallel, and another where the pair must coordinate to share one control. We found that a shared CAD control led to significantly better overall CAD quality than parallel CAD control. In addition, the shared control mode led to more complete and consistent CAD models, as well as the tendency for participants to follow instructions to correctly replicate features for the design task. As is predicted in the literature, a trade-off relationship (albeit weak) between quality and speed via the parallel collaboration was found. In contrast, the shared control mode shows no clear relationship between speed and quality. Collaborative CAD is increasingly seen as an appealing tool for modern product design teams. This study suggests that the benefits of this tool are not solely the effect of the tool itself, but result from the collaboration style of the designers using the tool.

The Research on Automatic Liquor Blending Systems Based on Wavelet Neural Networks

2013 ◽  
Vol 765-767 ◽  
pp. 1019-1022
Author(s):  
Lian Jun Hu ◽  
Xiao Hui Zeng ◽  
Hong Song ◽  
Qian Li
Keyword(s):  
Neural Networks ◽  
Computer Aided Design ◽  
Mathematic Model ◽  
Wavelet Neural Networks ◽  
Time Frequency ◽  
The Neural Network ◽  
Computer Aided ◽  
Time Frequency Localization ◽  
Aided Design ◽  
Self Learning

The blending of liquors is a key process in the production of liquors. According to time-frequency localization characteristics of the wavelet transform and advantages of the neural network such as ability to develop, fault-tolerance, self-adaptability, self-learning, and robustness, a mathematic model based on wavelet neural networks is proposed in liquor blending processes with the help of computer-aided design technologies, which makes liquor blending technologies more scientific.

Computation of Midsurface by Feature-Based Simplification–Abstraction–Decomposition

2016 ◽  
Vol 17 (1) ◽  
Author(s):  
Yogesh H. Kulkarni ◽  
Anil Sahasrabudhe ◽  
Mukund Kale
Keyword(s):  
Computer Aided Design ◽  
Real Life ◽  
Boundary Representation ◽  
Manual Task ◽  
Computer Aided ◽  
Cad Models ◽  
Feature Based ◽  
Feature Information ◽  
Plastic Parts ◽  
Aided Design

Computer-aided design (CAD) models of thin-walled solids such as sheet metal or plastic parts are often reduced dimensionally to their corresponding midsurfaces for quicker and fairly accurate results of computer-aided engineering (CAE) analysis. Computation of the midsurface is still a time-consuming and mostly, a manual task due to lack of robust and automated techniques. Most of the existing techniques work on the final shape (typically in the form of boundary representation, B-rep). Complex B-reps make it hard to detect subshapes for which the midsurface patches are computed and joined, forcing usage of hard-coded heuristic rules, developed on a case-by-case basis. Midsurface failures manifest in the form of gaps, overlaps, nonmimicking input model, etc., which can take hours or even days to correct. The research presented here proposes to address these problems by leveraging feature-information available in the modern CAD models, and by effectively using techniques like simplification, abstraction, and decomposition. In the proposed approach, first, the irrelevant features are identified and removed from the input FbCAD model to compute its simplified gross shape. Remaining features then undergo abstraction to transform into their corresponding generic Loft-equivalents, each having a profile and a guide curve. The model is then decomposed into cellular bodies and a graph is populated, with cellular bodies at the nodes and fully overlapping-surface-interfaces at the edges. The nodes are classified into midsurface-patch generating nodes (called “solid cells” or sCells) and interaction-resolving nodes (“interface cells” or iCells). In a sCell, a midsurface patch is generated either by offset or by sweeping the midcurve of the owner-Loft-feature's profile along with its guide curve. Midsurface patches are then connected in the iCells in a generic manner, thus resulting in a well-connected midsurface with minimum failures. Output midsurface is then validated topologically for correctness. At the end of this paper, real-life parts are used to demonstrate the efficacy of the proposed approach.

Sign in / Sign up

Export Citation Format

Share Document