Frank Weichert

Bio: Frank Weichert is an academic researcher from Technical University of Dortmund. The author has contributed to research in topics: Finite element method & Mesh generation. The author has an hindex of 15, co-authored 108 publications receiving 1823 citations. Previous affiliations of Frank Weichert include Leibniz Association & Fraunhofer Society.

Analysis of the Accuracy and Robustness of the Leap Motion Controller

Abstract: The Leap Motion Controller is a new device for hand gesture controlled user interfaces with declared sub-millimeter accuracy However, up to this point its capabilities in real environments have not been analyzed Therefore, this paper presents a first study of a Leap Motion Controller The main focus of attention is on the evaluation of the accuracy and repeatability For an appropriate evaluation, a novel experimental setup was developed making use of an industrial robot with a reference pen allowing a position accuracy of 02 mm Thereby, a deviation between a desired 3D position and the average measured positions below 02mmhas been obtained for static setups and of 12mmfor dynamic setups Using the conclusion of this analysis can improve the development of applications for the Leap Motion controller in the field of Human-Computer Interaction

SplineCNN: Fast Geometric Deep Learning with Continuous B-Spline Kernels

Abstract: We present Spline-based Convolutional Neural Networks (SplineCNNs), a variant of deep neural networks for irregular structured and geometric input, e.g., graphs or meshes. Our main contribution is a novel convolution operator based on B-splines, that makes the computation time independent from the kernel size due to the local support property of the B-spline basis functions. As a result, we obtain a generalization of the traditional CNN convolution operator by using continuous kernel functions parametrized by a fixed number of trainable weights. In contrast to related approaches that filter in the spectral domain, the proposed method aggregates features purely in the spatial domain. In addition, SplineCNN allows entire end-to-end training of deep architectures, using only the geometric structure as input, instead of handcrafted feature descriptors. For validation, we apply our method on tasks from the fields of image graph classification, shape correspondence and graph node classification, and show that it outperforms or pars state-of-the-art approaches while being significantly faster and having favorable properties like domain-independence. Our source code is available on GitHub1.

Evaluation of the leap motion controller as a new contact-free pointing device.

Abstract: This paper presents a Fitts' law-based analysis of the user's performance in selection tasks with the Leap Motion Controller compared with a standard mouse device. The Leap Motion Controller (LMC) is a new contact-free input system for gesture-based human-computer interaction with declared sub-millimeter accuracy. Up to this point, there has hardly been any systematic evaluation of this new system available. With an error rate of 7.8 % for the LMC and 2.8% for the mouse device, movement times twice as large as for a mouse device and high overall effort ratings, the Leap Motion Controller's performance as an input device for everyday generic computer pointing tasks is rather limited, at least with regard to the selection recognition provided by the LMC.

Review of Three-Dimensional Human-Computer Interaction with Focus on the Leap Motion Controller

Abstract: Modern hardware and software development has led to an evolution of user interfaces from command-line to natural user interfaces for virtual immersive environments. Gestures imitating real-world interaction tasks increasingly replace classical two-dimensional interfaces based on Windows/Icons/Menus/Pointers (WIMP) or touch metaphors. Thus, the purpose of this paper is to survey the state-of-the-art Human-Computer Interaction (HCI) techniques with a focus on the special field of three-dimensional interaction. This includes an overview of currently available interaction devices, their applications of usage and underlying methods for gesture design and recognition. Focus is on interfaces based on the Leap Motion Controller (LMC) and corresponding methods of gesture design and recognition. Further, a review of evaluation methods for the proposed natural user interfaces is given.

Hierarchical Inter-Message Passing for Learning on Molecular Graphs.

Abstract: We present a hierarchical neural message passing architecture for learning on molecular graphs Our model takes in two complementary graph representations: the raw molecular graph representation and its associated junction tree, where nodes represent meaningful clusters in the original graph, eg, rings or bridged compounds We then proceed to learn a molecule's representation by passing messages inside each graph, and exchange messages between the two representations using a coarse-to-fine and fine-to-coarse information flow Our method is able to overcome some of the restrictions known from classical GNNs, like detecting cycles, while still being very efficient to train We validate its performance on the ZINC dataset and datasets stemming from the MoleculeNet benchmark collection

A Comprehensive Survey on Graph Neural Networks

Abstract: Deep learning has revolutionized many machine learning tasks in recent years, ranging from image classification and video processing to speech recognition and natural language understanding. The data in these tasks are typically represented in the Euclidean space. However, there is an increasing number of applications, where data are generated from non-Euclidean domains and are represented as graphs with complex relationships and interdependency between objects. The complexity of graph data has imposed significant challenges on the existing machine learning algorithms. Recently, many studies on extending deep learning approaches for graph data have emerged. In this article, we provide a comprehensive overview of graph neural networks (GNNs) in data mining and machine learning fields. We propose a new taxonomy to divide the state-of-the-art GNNs into four categories, namely, recurrent GNNs, convolutional GNNs, graph autoencoders, and spatial–temporal GNNs. We further discuss the applications of GNNs across various domains and summarize the open-source codes, benchmark data sets, and model evaluation of GNNs. Finally, we propose potential research directions in this rapidly growing field.

Dynamic Graph CNN for Learning on Point Clouds

Abstract: Point clouds provide a flexible geometric representation suitable for countless applications in computer graphics; they also comprise the raw output of most 3D data acquisition devices. While hand-designed features on point clouds have long been proposed in graphics and vision, however, the recent overwhelming success of convolutional neural networks (CNNs) for image analysis suggests the value of adapting insight from CNN to the point cloud world. Point clouds inherently lack topological information, so designing a model to recover topology can enrich the representation power of point clouds. To this end, we propose a new neural network module dubbed EdgeConv suitable for CNN-based high-level tasks on point clouds, including classification and segmentation. EdgeConv acts on graphs dynamically computed in each layer of the network. It is differentiable and can be plugged into existing architectures. Compared to existing modules operating in extrinsic space or treating each point independently, EdgeConv has several appealing properties: It incorporates local neighborhood information; it can be stacked applied to learn global shape properties; and in multi-layer systems affinity in feature space captures semantic characteristics over potentially long distances in the original embedding. We show the performance of our model on standard benchmarks, including ModelNet40, ShapeNetPart, and S3DIS.

Fast Graph Representation Learning with PyTorch Geometric

Abstract: We introduce PyTorch Geometric, a library for deep learning on irregularly structured input data such as graphs, point clouds and manifolds, built upon PyTorch. In addition to general graph data structures and processing methods, it contains a variety of recently published methods from the domains of relational learning and 3D data processing. PyTorch Geometric achieves high data throughput by leveraging sparse GPU acceleration, by providing dedicated CUDA kernels and by introducing efficient mini-batch handling for input examples of different size. In this work, we present the library in detail and perform a comprehensive comparative study of the implemented methods in homogeneous evaluation scenarios.

Hierarchical Graph Representation Learning with Differentiable Pooling

Abstract: Recently, graph neural networks (GNNs) have revolutionized the field of graph representation learning through effectively learned node embeddings, and achieved state-of-the-art results in tasks such as node classification and link prediction. However, current GNN methods are inherently flat and do not learn hierarchical representations of graphs---a limitation that is especially problematic for the task of graph classification, where the goal is to predict the label associated with an entire graph. Here we propose DiffPool, a differentiable graph pooling module that can generate hierarchical representations of graphs and can be combined with various graph neural network architectures in an end-to-end fashion. DiffPool learns a differentiable soft cluster assignment for nodes at each layer of a deep GNN, mapping nodes to a set of clusters, which then form the coarsened input for the next GNN layer. Our experimental results show that combining existing GNN methods with DiffPool yields an average improvement of 5-10% accuracy on graph classification benchmarks, compared to all existing pooling approaches, achieving a new state-of-the-art on four out of five benchmark data sets.

Dynamic Graph CNN for Learning on Point Clouds

Abstract: Point clouds provide a flexible geometric representation suitable for countless applications in computer graphics; they also comprise the raw output of most 3D data acquisition devices. While hand-designed features on point clouds have long been proposed in graphics and vision, however, the recent overwhelming success of convolutional neural networks (CNNs) for image analysis suggests the value of adapting insight from CNN to the point cloud world. Point clouds inherently lack topological information so designing a model to recover topology can enrich the representation power of point clouds. To this end, we propose a new neural network module dubbed EdgeConv suitable for CNN-based high-level tasks on point clouds including classification and segmentation. EdgeConv acts on graphs dynamically computed in each layer of the network. It is differentiable and can be plugged into existing architectures. Compared to existing modules operating in extrinsic space or treating each point independently, EdgeConv has several appealing properties: It incorporates local neighborhood information; it can be stacked applied to learn global shape properties; and in multi-layer systems affinity in feature space captures semantic characteristics over potentially long distances in the original embedding. We show the performance of our model on standard benchmarks including ModelNet40, ShapeNetPart, and S3DIS.