Showing 1–7 of 7 results for author: Weichert, F

Visual Trajectory Prediction of Vessels for Inland Navigation

Authors: Alexander Puzicha, Konstantin Wüstefeld, Kathrin Wilms, Frank Weichert

Abstract: The future of inland navigation increasingly relies on autonomous systems and remote operations, emphasizing the need for accurate vessel trajectory prediction. This study addresses the challenges of video-based vessel tracking and prediction by integrating advanced object detection methods, Kalman filters, and spline-based interpolation. However, existing detection systems often misclassify objec… ▽ More The future of inland navigation increasingly relies on autonomous systems and remote operations, emphasizing the need for accurate vessel trajectory prediction. This study addresses the challenges of video-based vessel tracking and prediction by integrating advanced object detection methods, Kalman filters, and spline-based interpolation. However, existing detection systems often misclassify objects in inland waterways due to complex surroundings. A comparative evaluation of tracking algorithms, including BoT-SORT, Deep OC-SORT, and ByeTrack, highlights the robustness of the Kalman filter in providing smoothed trajectories. Experimental results from diverse scenarios demonstrate improved accuracy in predicting vessel movements, which is essential for collision avoidance and situational awareness. The findings underline the necessity of customized datasets and models for inland navigation. Future work will expand the datasets and incorporate vessel classification to refine predictions, supporting both autonomous systems and human operators in complex environments. △ Less

Submitted 1 May, 2025; originally announced May 2025.

Definition-independent Formalization of Soundscapes: Towards a Formal Methodology

Authors: Mikel D. Jedrusiak, Thomas Harweg, Timo Haselhoff, Bryce T. Lawrence, Susanne Moebus, Frank Weichert

Abstract: Soundscapes have been studied by researchers from various disciplines, each with different perspectives, goals, approaches, and terminologies. Accordingly, depending on the field, the concept of a soundscape's components changes, consequently changing the basic definition. This results in complicating interdisciplinary communication and comparison of results. Especially when soundscape-unrelated r… ▽ More Soundscapes have been studied by researchers from various disciplines, each with different perspectives, goals, approaches, and terminologies. Accordingly, depending on the field, the concept of a soundscape's components changes, consequently changing the basic definition. This results in complicating interdisciplinary communication and comparison of results. Especially when soundscape-unrelated research areas are involved. For this reason, we present a potential formalization that is independent of the underlying soundscape definition, with the goal of being able to capture the heterogeneous structure of the data as well as the different ideologies in one model. In an exemplary analysis of frequency correlation matrices for land use type detection as an alternative to features like MFCCs, we show a practical application of our presented formalization. △ Less

Submitted 20 October, 2023; originally announced October 2023.

GNNAutoScale: Scalable and Expressive Graph Neural Networks via Historical Embeddings

Authors: Matthias Fey, Jan E. Lenssen, Frank Weichert, Jure Leskovec

Abstract: We present GNNAutoScale (GAS), a framework for scaling arbitrary message-passing GNNs to large graphs. GAS prunes entire sub-trees of the computation graph by utilizing historical embeddings from prior training iterations, leading to constant GPU memory consumption in respect to input node size without dropping any data. While existing solutions weaken the expressive power of message passing due t… ▽ More We present GNNAutoScale (GAS), a framework for scaling arbitrary message-passing GNNs to large graphs. GAS prunes entire sub-trees of the computation graph by utilizing historical embeddings from prior training iterations, leading to constant GPU memory consumption in respect to input node size without dropping any data. While existing solutions weaken the expressive power of message passing due to sub-sampling of edges or non-trainable propagations, our approach is provably able to maintain the expressive power of the original GNN. We achieve this by providing approximation error bounds of historical embeddings and show how to tighten them in practice. Empirically, we show that the practical realization of our framework, PyGAS, an easy-to-use extension for PyTorch Geometric, is both fast and memory-efficient, learns expressive node representations, closely resembles the performance of their non-scaling counterparts, and reaches state-of-the-art performance on large-scale graphs. △ Less

Submitted 10 June, 2021; originally announced June 2021.

Comments: Published as a conference paper at ICML 2021

Hierarchical Inter-Message Passing for Learning on Molecular Graphs

Authors: Matthias Fey, Jan-Gin Yuen, Frank Weichert

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, e.g., rings or bridged compounds. We then proceed to learn a molecule's representation by passing messages… ▽ More 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, e.g., 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. △ Less

Submitted 22 June, 2020; originally announced June 2020.

Comments: Published at the Graph Representation Learning and Beyond (GRL+) Workhop at ICML 2020

Adversarial Generation of Continuous Implicit Shape Representations

Authors: Marian Kleineberg, Matthias Fey, Frank Weichert

Abstract: This work presents a generative adversarial architecture for generating three-dimensional shapes based on signed distance representations. While the deep generation of shapes has been mostly tackled by voxel and surface point cloud approaches, our generator learns to approximate the signed distance for any point in space given prior latent information. Although structurally similar to generative p… ▽ More This work presents a generative adversarial architecture for generating three-dimensional shapes based on signed distance representations. While the deep generation of shapes has been mostly tackled by voxel and surface point cloud approaches, our generator learns to approximate the signed distance for any point in space given prior latent information. Although structurally similar to generative point cloud approaches, this formulation can be evaluated with arbitrary point density during inference, leading to fine-grained details in generated outputs. Furthermore, we study the effects of using either progressively growing voxel- or point-processing networks as discriminators, and propose a refinement scheme to strengthen the generator's capabilities in modeling the zero iso-surface decision boundary of shapes. We train our approach on the ShapeNet benchmark dataset and validate, both quantitatively and qualitatively, its performance in generating realistic 3D shapes. △ Less

Submitted 26 March, 2020; v1 submitted 2 February, 2020; originally announced February 2020.

Comments: Published in Eurographics 2020 - Short Papers

Recognizing Cuneiform Signs Using Graph Based Methods

Authors: Nils M. Kriege, Matthias Fey, Denis Fisseler, Petra Mutzel, Frank Weichert

Abstract: The cuneiform script constitutes one of the earliest systems of writing and is realized by wedge-shaped marks on clay tablets. A tremendous number of cuneiform tablets have already been discovered and are incrementally digitalized and made available to automated processing. As reading cuneiform script is still a manual task, we address the real-world application of recognizing cuneiform signs by t… ▽ More The cuneiform script constitutes one of the earliest systems of writing and is realized by wedge-shaped marks on clay tablets. A tremendous number of cuneiform tablets have already been discovered and are incrementally digitalized and made available to automated processing. As reading cuneiform script is still a manual task, we address the real-world application of recognizing cuneiform signs by two graph based methods with complementary runtime characteristics. We present a graph model for cuneiform signs together with a tailored distance measure based on the concept of the graph edit distance. We propose efficient heuristics for its computation and demonstrate its effectiveness in classification tasks experimentally. To this end, the distance measure is used to implement a nearest neighbor classifier leading to a high computational cost for the prediction phase with increasing training set size. In order to overcome this issue, we propose to use CNNs adapted to graphs as an alternative approach shifting the computational cost to the training phase. We demonstrate the practicability of both approaches in an extensive experimental comparison regarding runtime and prediction accuracy. Although currently available annotated real-world data is still limited, we obtain a high accuracy using CNNs, in particular, when the training set is enriched by augmented examples. △ Less

Submitted 9 March, 2018; v1 submitted 16 February, 2018; originally announced February 2018.

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

Authors: Matthias Fey, Jan Eric Lenssen, Frank Weichert, Heinrich Müller

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… ▽ More 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. △ Less

Submitted 23 May, 2018; v1 submitted 24 November, 2017; originally announced November 2017.

Comments: Presented at CVPR 2018