Browsing by Author "Sprenger, Janis"

Now showing 1 - 3 of 3
  • Thumbnail Image
    Item
    Fine-Grained Semantic Segmentation of Motion Capture Data using Dilated Temporal Fully-Convolutional Networks
    (The Eurographics Association, 2019) Cheema, Noshaba; hosseini, somayeh; Sprenger, Janis; Herrmann, Erik; Du, Han; Fischer, Klaus; Slusallek, Philipp; Cignoni, Paolo and Miguel, Eder
    Human motion capture data has been widely used in data-driven character animation. In order to generate realistic, naturallooking motions, most data-driven approaches require considerable efforts of pre-processing, including motion segmentation and annotation. Existing (semi-) automatic solutions either require hand-crafted features for motion segmentation or do not produce the semantic annotations required for motion synthesis and building large-scale motion databases. In addition, human labeled annotation data suffers from inter- and intra-labeler inconsistencies by design. We propose a semi-automatic framework for semantic segmentation of motion capture data based on supervised machine learning techniques. It first transforms a motion capture sequence into a ''motion image'' and applies a convolutional neural network for image segmentation. Dilated temporal convolutions enable the extraction of temporal information from a large receptive field. Our model outperforms two state-of-the-art models for action segmentation, as well as a popular network for sequence modeling. Most of all, our method is very robust under noisy and inaccurate training labels and thus can handle human errors during the labeling process.
  • No Thumbnail Available
    Item
    Motion Data and Model Management for Applied Statistical Motion Synthesis
    (The Eurographics Association, 2019) Herrmann, Erik; Du, Han; Antakli, André; Rubinstein, Dmitri; Schubotz, René; Sprenger, Janis; Hosseini, Somayeh; Cheema, Noshaba; Zinnikus, Ingo; Manns, Martin; Fischer, Klaus; Slusallek, Philipp; Agus, Marco and Corsini, Massimiliano and Pintus, Ruggero
    Machine learning based motion modelling methods such as statistical modelling require a large amount of input data. In practice, the management of the data can become a problem in itself for artists who want to control the quality of the motion models. As a solution to this problem, we present a motion data and model management system and integrate it with a statistical motion modelling pipeline. The system is based on a data storage server with a REST interface that enables the efficient storage of different versions of motion data and models. The database system is combined with a motion preprocessing tool that provides functions for batch editing, retargeting and annotation of the data. For the application of the motion models in a game engine, the framework provides a stateful motion synthesis server that can load the models directly from the data storage server. Additionally, the framework makes use of a Kubernetes compute cluster to execute time consuming processes such as the preprocessing and modelling of the data. The system is evaluated in a use case for the simulation of manual assembly workers.
  • Thumbnail Image
    Item
    Stylistic Locomotion Modeling with Conditional Variational Autoencoder
    (The Eurographics Association, 2019) Du, Han; Herrmann, Erik; Sprenger, Janis; Cheema, Noshaba; hosseini, somayeh; Fischer, Klaus; Slusallek, Philipp; Cignoni, Paolo and Miguel, Eder
    We propose a novel approach to create generative models for distinctive stylistic locomotion synthesis. The approach is inspired by the observation that human styles can be easily distinguished from a few examples. However, learning a generative model for natural human motions which display huge amounts of variations and randomness would require a lot of training data. Furthermore, it would require considerable efforts to create such a large motion database for each style. We propose a generative model to combine the large variation in a neutral motion database and style information from a limited number of examples. We formulate the stylistic motion modeling task as a conditional distribution learning problem. Style transfer is implicitly applied during the model learning process. A conditional variational autoencoder (CVAE) is applied to learn the distribution and stylistic examples are used as constraints. We demonstrate that our approach can generate any number of natural-looking human motions with a similar style to the target given a few style examples and a neutral motion database.