Header logo is


2017


no image
Phase Estimation for Fast Action Recognition and Trajectory Generation in Human-Robot Collaboration

Maeda, G., Ewerton, M., Neumann, G., Lioutikov, R., Peters, J.

International Journal of Robotics Research, 36(13-14):1579-1594, 2017, Special Issue on the Seventeenth International Symposium on Robotics Research (article)

ei

DOI Project Page [BibTex]

2017


DOI Project Page [BibTex]


no image
Causal models for decision making via integrative inference

Geiger, P.

University of Stuttgart, Germany, 2017 (phdthesis)

ei

[BibTex]

[BibTex]


no image
A Phase-coded Aperture Camera with Programmable Optics

Chen, J., Hirsch, M., Heintzmann, R., Eberhardt, B., Lensch, H. P. A.

Electronic Imaging, 2017(17):70-75, 2017 (article)

ei

DOI [BibTex]

DOI [BibTex]


Capturing Hand-Object Interaction and Reconstruction of Manipulated Objects
Capturing Hand-Object Interaction and Reconstruction of Manipulated Objects

Tzionas, D.

University of Bonn, 2017 (phdthesis)

Abstract
Hand motion capture with an RGB-D sensor gained recently a lot of research attention, however, even most recent approaches focus on the case of a single isolated hand. We focus instead on hands that interact with other hands or with a rigid or articulated object. Our framework successfully captures motion in such scenarios by combining a generative model with discriminatively trained salient points, collision detection and physics simulation to achieve a low tracking error with physically plausible poses. All components are unified in a single objective function that can be optimized with standard optimization techniques. We initially assume a-priori knowledge of the object's shape and skeleton. In case of unknown object shape there are existing 3d reconstruction methods that capitalize on distinctive geometric or texture features. These methods though fail for textureless and highly symmetric objects like household articles, mechanical parts or toys. We show that extracting 3d hand motion for in-hand scanning effectively facilitates the reconstruction of such objects and we fuse the rich additional information of hands into a 3d reconstruction pipeline. Finally, although shape reconstruction is enough for rigid objects, there is a lack of tools that build rigged models of articulated objects that deform realistically using RGB-D data. We propose a method that creates a fully rigged model consisting of a watertight mesh, embedded skeleton and skinning weights by employing a combination of deformable mesh tracking, motion segmentation based on spectral clustering and skeletonization based on mean curvature flow.

ps

Thesis link (url) Project Page [BibTex]


no image
On Maximum Entropy and Inference

Gresele, L., Marsili, M.

Entropy, 19(12):article no. 642, 2017 (article)

ei

link (url) [BibTex]

link (url) [BibTex]


no image
Towards Engagement Models that Consider Individual Factors in HRI: On the Relation of Extroversion and Negative Attitude Towards Robots to Gaze and Speech During a Human-Robot Assembly Task

Ivaldi, S., Lefort, S., Peters, J., Chetouani, M., Provasi, J., Zibetti, E.

International Journal of Social Robotics, 9(1):63-86, 2017 (article)

ei

DOI [BibTex]

DOI [BibTex]


no image
Non-parametric Policy Search with Limited Information Loss

van Hoof, H., Neumann, G., Peters, J.

Journal of Machine Learning Research , 18(73):1-46, 2017 (article)

ei

link (url) Project Page [BibTex]

link (url) Project Page [BibTex]


no image
Stability of Controllers for Gaussian Process Dynamics

Vinogradska, J., Bischoff, B., Nguyen-Tuong, D., Peters, J.

Journal of Machine Learning Research, 18(100):1-37, 2017 (article)

ei

link (url) Project Page [BibTex]

link (url) Project Page [BibTex]


no image
Learning Optimal Configurations for Modeling Frowning by Transcranial Electrical Stimulation

Sücker, K.

Graduate School of Neural Information Processing, Eberhard Karls Universität Tübingen, Germany, 2017 (mastersthesis)

ei

[BibTex]

[BibTex]


no image
SUV-quantification of physiological lung tissue in an integrated PET/MR-system: Impact of lung density and bone tissue

Seith, F., Schmidt, H., Gatidis, S., Bezrukov, I., Schraml, C., Pfannenberg, C., la Fougère, C., Nikolaou, K., Schwenzer, N.

PLOS ONE, 12(5):1-13, 2017 (article)

ei

DOI [BibTex]

DOI [BibTex]


no image
Momentum-Centered Control of Contact Interactions

Righetti, L., Herzog, A.

In Geometric and Numerical Foundations of Movements, 117, pages: 339-359, Springer Tracts in Advanced Robotics, Springer, Cham, 2017 (incollection)

mg

link (url) [BibTex]

link (url) [BibTex]


no image
Pattern Generation for Walking on Slippery Terrains

Khadiv, M., Moosavian, S. A. A., Herzog, A., Righetti, L.

In 2017 5th International Conference on Robotics and Mechatronics (ICROM), Iran, August 2017 (inproceedings)

Abstract
In this paper, we extend state of the art Model Predictive Control (MPC) approaches to generate safe bipedal walking on slippery surfaces. In this setting, we formulate walking as a trade off between realizing a desired walking velocity and preserving robust foot-ground contact. Exploiting this for- mulation inside MPC, we show that safe walking on various flat terrains can be achieved by compromising three main attributes, i. e. walking velocity tracking, the Zero Moment Point (ZMP) modulation, and the Required Coefficient of Friction (RCoF) regulation. Simulation results show that increasing the walking velocity increases the possibility of slippage, while reducing the slippage possibility conflicts with reducing the tip-over possibility of the contact and vice versa.

mg

link (url) [BibTex]

link (url) [BibTex]

2016


no image
Consistent Kernel Mean Estimation for Functions of Random Variables

Simon-Gabriel*, C. J., Ścibior*, A., Tolstikhin, I., Schölkopf, B.

Advances in Neural Information Processing Systems 29, pages: 1732-1740, (Editors: D. D. Lee, M. Sugiyama, U. V. Luxburg, I. Guyon, and R. Garnett), Curran Associates, Inc., 30th Annual Conference on Neural Information Processing Systems, December 2016, *joint first authors (conference)

ei

link (url) Project Page Project Page Project Page [BibTex]

2016


link (url) Project Page Project Page Project Page [BibTex]


no image
Understanding Probabilistic Sparse Gaussian Process Approximations

Bauer, M., van der Wilk, M., Rasmussen, C. E.

Advances in Neural Information Processing Systems 29, pages: 1533-1541, (Editors: D. D. Lee, M. Sugiyama, U. V. Luxburg, I. Guyon, and R. Garnett), Curran Associates, Inc., 30th Annual Conference on Neural Information Processing Systems, December 2016 (conference)

ei

link (url) Project Page [BibTex]

link (url) Project Page [BibTex]


no image
Minimax Estimation of Maximum Mean Discrepancy with Radial Kernels

Tolstikhin, I., Sriperumbudur, B. K., Schölkopf, B.

Advances in Neural Information Processing Systems 29, pages: 1930-1938, (Editors: D. D. Lee, M. Sugiyama, U. V. Luxburg, I. Guyon, and R. Garnett), Curran Associates, Inc., 30th Annual Conference on Neural Information Processing Systems, December 2016 (conference)

ei

link (url) Project Page [BibTex]

link (url) Project Page [BibTex]


no image
Local-utopia Policy Selection for Multi-objective Reinforcement Learning

Parisi, S., Blank, A., Viernickel, T., Peters, J.

In IEEE Symposium on Adaptive Dynamic Programming and Reinforcement Learning (ADPRL), pages: 1-7, IEEE, December 2016 (inproceedings)

ei

DOI [BibTex]

DOI [BibTex]


Skinned multi-person linear model
Skinned multi-person linear model

Black, M.J., Loper, M., Mahmood, N., Pons-Moll, G., Romero, J.

December 2016, Application PCT/EP2016/064610 (misc)

Abstract
The invention comprises a learned model of human body shape and pose dependent shape variation that is more accurate than previous models and is compatible with existing graphics pipelines. Our Skinned Multi-Person Linear model (SMPL) is a skinned vertex based model that accurately represents a wide variety of body shapes in natural human poses. The parameters of the model are learned from data including the rest pose template, blend weights, pose-dependent blend shapes, identity- dependent blend shapes, and a regressor from vertices to joint locations. Unlike previous models, the pose-dependent blend shapes are a linear function of the elements of the pose rotation matrices. This simple formulation enables training the entire model from a relatively large number of aligned 3D meshes of different people in different poses. The invention quantitatively evaluates variants of SMPL using linear or dual- quaternion blend skinning and show that both are more accurate than a Blend SCAPE model trained on the same data. In a further embodiment, the invention realistically models dynamic soft-tissue deformations. Because it is based on blend skinning, SMPL is compatible with existing rendering engines and we make it available for research purposes.

ps

Google Patents [BibTex]

Google Patents [BibTex]


no image
Lifelong Learning with Weighted Majority Votes

Pentina, A., Urner, R.

Advances in Neural Information Processing Systems 29, pages: 3612-3620, (Editors: D. D. Lee, M. Sugiyama, U. V. Luxburg, I. Guyon, and R. Garnett), Curran Associates, Inc., 30th Annual Conference on Neural Information Processing Systems, December 2016 (conference)

ei

link (url) Project Page [BibTex]

link (url) Project Page [BibTex]


no image
Active Nearest-Neighbor Learning in Metric Spaces

Kontorovich, A., Sabato, S., Urner, R.

Advances in Neural Information Processing Systems 29, pages: 856-864, (Editors: D. D. Lee, M. Sugiyama, U. V. Luxburg, I. Guyon, and R. Garnett), Curran Associates, Inc., 30th Annual Conference on Neural Information Processing Systems, December 2016 (conference)

ei

link (url) Project Page [BibTex]

link (url) Project Page [BibTex]


no image
Catching heuristics are optimal control policies

Belousov, B., Neumann, G., Rothkopf, C., Peters, J.

Advances in Neural Information Processing Systems 29, pages: 1426-1434, (Editors: D. D. Lee, M. Sugiyama, U. V. Luxburg, I. Guyon, and R. Garnett), Curran Associates, Inc., 30th Annual Conference on Neural Information Processing Systems, December 2016 (conference)

ei

link (url) Project Page [BibTex]

link (url) Project Page [BibTex]


no image
Incremental Imitation Learning of Context-Dependent Motor Skills

Ewerton, M., Maeda, G., Kollegger, G., Wiemeyer, J., Peters, J.

IEEE-RAS 16th International Conference on Humanoid Robots (Humanoids), pages: 351-358, IEEE, November 2016 (conference)

ei

DOI Project Page [BibTex]

DOI Project Page [BibTex]


no image
Using Probabilistic Movement Primitives for Striking Movements

Gomez-Gonzalez, S., Neumann, G., Schölkopf, B., Peters, J.

16th IEEE-RAS International Conference on Humanoid Robots (Humanoids), pages: 502-508, November 2016 (conference)

am ei

link (url) DOI Project Page [BibTex]

link (url) DOI Project Page [BibTex]


no image
Demonstration Based Trajectory Optimization for Generalizable Robot Motions

Koert, D., Maeda, G., Lioutikov, R., Neumann, G., Peters, J.

IEEE-RAS 16th International Conference on Humanoid Robots (Humanoids), pages: 351-358, IEEE, November 2016 (conference)

ei

DOI Project Page [BibTex]

DOI Project Page [BibTex]


Jointly Learning Trajectory Generation and Hitting Point Prediction in Robot Table Tennis
Jointly Learning Trajectory Generation and Hitting Point Prediction in Robot Table Tennis

Huang, Y., Büchler, D., Koc, O., Schölkopf, B., Peters, J.

16th IEEE-RAS International Conference on Humanoid Robots (Humanoids), pages: 650-655, November 2016 (conference)

am ei

final link (url) DOI Project Page [BibTex]

final link (url) DOI Project Page [BibTex]


no image
Deep Spiking Networks for Model-based Planning in Humanoids

Tanneberg, D., Paraschos, A., Peters, J., Rueckert, E.

IEEE-RAS 16th International Conference on Humanoid Robots (Humanoids), pages: 656-661, IEEE, November 2016 (conference)

ei

DOI Project Page [BibTex]

DOI Project Page [BibTex]


no image
Anticipative Interaction Primitives for Human-Robot Collaboration

Maeda, G., Maloo, A., Ewerton, M., Lioutikov, R., Peters, J.

AAAI Fall Symposium Series. Shared Autonomy in Research and Practice, pages: 325-330, November 2016 (conference)

ei

link (url) [BibTex]

link (url) [BibTex]


Creating body shapes from verbal descriptions by linking similarity spaces
Creating body shapes from verbal descriptions by linking similarity spaces

Hill, M. Q., Streuber, S., Hahn, C. A., Black, M. J., O’Toole, A. J.

Psychological Science, 27(11):1486-1497, November 2016, (article)

Abstract
Brief verbal descriptions of bodies (e.g. curvy, long-legged) can elicit vivid mental images. The ease with which we create these mental images belies the complexity of three-dimensional body shapes. We explored the relationship between body shapes and body descriptions and show that a small number of words can be used to generate categorically accurate representations of three-dimensional bodies. The dimensions of body shape variation that emerged in a language-based similarity space were related to major dimensions of variation computed directly from three-dimensional laser scans of 2094 bodies. This allowed us to generate three-dimensional models of people in the shape space using only their coordinates on analogous dimensions in the language-based description space. Human descriptions of photographed bodies and their corresponding models matched closely. The natural mapping between the spaces illustrates the role of language as a concise code for body shape, capturing perceptually salient global and local body features.

ps

pdf [BibTex]

pdf [BibTex]


no image
Unifying distillation and privileged information

Lopez-Paz, D., Schölkopf, B., Bottou, L., Vapnik, V.

International Conference on Learning Representations (ICLR), November 2016 (conference)

ei

Arxiv Project Page [BibTex]

Arxiv Project Page [BibTex]


no image
Learning High-Order Filters for Efficient Blind Deconvolution of Document Photographs

Xiao, L., Wang, J., Heidrich, W., Hirsch, M.

Computer Vision - ECCV 2016, Lecture Notes in Computer Science, LNCS 9907, Part III, pages: 734-749, (Editors: Bastian Leibe, Jiri Matas, Nicu Sebe and Max Welling), Springer, October 2016 (conference)

ei

DOI [BibTex]

DOI [BibTex]


no image
Adaptive Training Strategies for BCIs

Sharma, D., Tanneberg, D., Grosse-Wentrup, M., Peters, J., Rueckert, E.

Cybathlon Symposium, October 2016 (conference)

ei

link (url) [BibTex]

link (url) [BibTex]


no image
Experiments with Hierarchical Reinforcement Learning of Multiple Grasping Policies

Osa, T., Peters, J., Neumann, G.

International Symposium on Experimental Robotics (ISER), 1, pages: 160-172, Springer Proceedings in Advanced Robotics, (Editors: Dana Kulic, Yoshihiko Nakamura, Oussama Khatib and Gentiane Venture), Springer, October 2016 (conference)

ei

DOI [BibTex]

DOI [BibTex]


no image
Stable Reinforcement Learning with Autoencoders for Tactile and Visual Data

van Hoof, H., Chen, N., Karl, M., van der Smagt, P., Peters, J.

Proceedings of the IEEE/RSJ Conference on Intelligent Robots and Systems (IROS), pages: 3928-3934, IEEE, October 2016 (conference)

ei

DOI Project Page [BibTex]

DOI Project Page [BibTex]


Keep it {SMPL}: Automatic Estimation of {3D} Human Pose and Shape from a Single Image
Keep it SMPL: Automatic Estimation of 3D Human Pose and Shape from a Single Image

Bogo, F., Kanazawa, A., Lassner, C., Gehler, P., Romero, J., Black, M. J.

In Computer Vision – ECCV 2016, pages: 561-578, Lecture Notes in Computer Science, Springer International Publishing, 14th European Conference on Computer Vision, October 2016 (inproceedings)

Abstract
We describe the first method to automatically estimate the 3D pose of the human body as well as its 3D shape from a single unconstrained image. We estimate a full 3D mesh and show that 2D joints alone carry a surprising amount of information about body shape. The problem is challenging because of the complexity of the human body, articulation, occlusion, clothing, lighting, and the inherent ambiguity in inferring 3D from 2D. To solve this, we fi rst use a recently published CNN-based method, DeepCut, to predict (bottom-up) the 2D body joint locations. We then fit (top-down) a recently published statistical body shape model, called SMPL, to the 2D joints. We do so by minimizing an objective function that penalizes the error between the projected 3D model joints and detected 2D joints. Because SMPL captures correlations in human shape across the population, we are able to robustly fi t it to very little data. We further leverage the 3D model to prevent solutions that cause interpenetration. We evaluate our method, SMPLify, on the Leeds Sports, HumanEva, and Human3.6M datasets, showing superior pose accuracy with respect to the state of the art.

ps

pdf Video Sup Mat video Code Project Project Page [BibTex]

pdf Video Sup Mat video Code Project Project Page [BibTex]


no image
A New Trajectory Generation Framework in Robotic Table Tennis

Koc, O., Maeda, G., Peters, J.

Proceedings of the IEEE/RSJ Conference on Intelligent Robots and Systems (IROS), pages: 3750-3756, October 2016 (conference)

am ei

link (url) DOI [BibTex]

link (url) DOI [BibTex]


Superpixel Convolutional Networks using Bilateral Inceptions
Superpixel Convolutional Networks using Bilateral Inceptions

Gadde, R., Jampani, V., Kiefel, M., Kappler, D., Gehler, P.

In European Conference on Computer Vision (ECCV), Lecture Notes in Computer Science, Springer, 14th European Conference on Computer Vision, October 2016 (inproceedings)

Abstract
In this paper we propose a CNN architecture for semantic image segmentation. We introduce a new “bilateral inception” module that can be inserted in existing CNN architectures and performs bilateral filtering, at multiple feature-scales, between superpixels in an image. The feature spaces for bilateral filtering and other parameters of the module are learned end-to-end using standard backpropagation techniques. The bilateral inception module addresses two issues that arise with general CNN segmentation architectures. First, this module propagates information between (super) pixels while respecting image edges, thus using the structured information of the problem for improved results. Second, the layer recovers a full resolution segmentation result from the lower resolution solution of a CNN. In the experiments, we modify several existing CNN architectures by inserting our inception modules between the last CNN (1 × 1 convolution) layers. Empirical results on three different datasets show reliable improvements not only in comparison to the baseline networks, but also in comparison to several dense-pixel prediction techniques such as CRFs, while being competitive in time.

am ps

pdf supplementary poster Project Page Project Page [BibTex]

pdf supplementary poster Project Page Project Page [BibTex]


no image
Probabilistic Decomposition of Sequential Force Interaction Tasks into Movement Primitives

Manschitz, S., Gienger, M., Kober, J., Peters, J.

IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS), pages: 3920-3927, IEEE, October 2016 (conference)

ei

DOI Project Page [BibTex]

DOI Project Page [BibTex]


Barrista - Caffe Well-Served
Barrista - Caffe Well-Served

Lassner, C., Kappler, D., Kiefel, M., Gehler, P.

In ACM Multimedia Open Source Software Competition, ACM OSSC16, October 2016 (inproceedings)

Abstract
The caffe framework is one of the leading deep learning toolboxes in the machine learning and computer vision community. While it offers efficiency and configurability, it falls short of a full interface to Python. With increasingly involved procedures for training deep networks and reaching depths of hundreds of layers, creating configuration files and keeping them consistent becomes an error prone process. We introduce the barrista framework, offering full, pythonic control over caffe. It separates responsibilities and offers code to solve frequently occurring tasks for pre-processing, training and model inspection. It is compatible to all caffe versions since mid 2015 and can import and export .prototxt files. Examples are included, e.g., a deep residual network implemented in only 172 lines (for arbitrary depths), comparing to 2320 lines in the official implementation for the equivalent model.

am ps

pdf link (url) DOI Project Page [BibTex]

pdf link (url) DOI Project Page [BibTex]


no image
Multi-task logistic regression in brain-computer interfaces

Fiebig, K., Jayaram, V., Peters, J., Grosse-Wentrup, M.

6th Workshop on Brain-Machine Interface Systems at IEEE International Conference on Systems, Man, and Cybernetics (SMC 2016), pages: 002307-002312, IEEE, October 2016 (conference)

ei

link (url) DOI [BibTex]

link (url) DOI [BibTex]


no image
Active Tactile Object Exploration with Gaussian Processes

Yi, Z., Calandra, R., Veiga, F., van Hoof, H., Hermans, T., Zhang, Y., Peters, J.

IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS), pages: 4925-4930, IEEE, October 2016 (conference)

ei

DOI Project Page [BibTex]

DOI Project Page [BibTex]


no image
On Version Space Compression

Ben-David, S., Urner, R.

Algorithmic Learning Theory - 27th International Conference (ALT), 9925, pages: 50-64, Lecture Notes in Computer Science, (Editors: Ortner, R., Simon, H. U., and Zilles, S.), September 2016 (conference)

ei

DOI Project Page [BibTex]

DOI Project Page [BibTex]


no image
Contextual Policy Search for Linear and Nonlinear Generalization of a Humanoid Walking Controller

Abdolmaleki, A., Lau, N., Reis, L., Peters, J., Neumann, G.

Journal of Intelligent & Robotic Systems, 83(3-4):393-408, (Editors: Luis Almeida, Lino Marques ), September 2016, Special Issue: Autonomous Robot Systems (article)

ei

DOI [BibTex]

DOI [BibTex]


no image
Learning Probabilistic Features from EMG Data for Predicting Knee Abnormalities

Kohlschuetter, J., Peters, J., Rueckert, E.

XIV Mediterranean Conference on Medical and Biological Engineering and Computing (MEDICON), pages: 668-672, (Editors: Kyriacou, E., Christofides, S., and Pattichis, C. S.), September 2016 (conference)

ei

DOI [BibTex]

DOI [BibTex]


no image
Planning with Information-Processing Constraints and Model Uncertainty in Markov Decision Processes

Grau-Moya, J, Leibfried, F, Genewein, T, Braun, DA

Machine Learning and Knowledge Discovery in Databases, pages: 475-491, Lecture Notes in Computer Science; 9852, Springer, Cham, Switzerland, European Conference on Machine Learning and Principles and Practice of Knowledge Discovery (ECML PKDD), September 2016 (conference)

Abstract
Information-theoretic principles for learning and acting have been proposed to solve particular classes of Markov Decision Problems. Mathematically, such approaches are governed by a variational free energy principle and allow solving MDP planning problems with information-processing constraints expressed in terms of a Kullback-Leibler divergence with respect to a reference distribution. Here we consider a generalization of such MDP planners by taking model uncertainty into account. As model uncertainty can also be formalized as an information-processing constraint, we can derive a unified solution from a single generalized variational principle. We provide a generalized value iteration scheme together with a convergence proof. As limit cases, this generalized scheme includes standard value iteration with a known model, Bayesian MDP planning, and robust planning. We demonstrate the benefits of this approach in a grid world simulation.

ei

DOI [BibTex]

DOI [BibTex]


Depth Estimation Through a Generative Model of Light Field Synthesis
Depth Estimation Through a Generative Model of Light Field Synthesis

Sajjadi, M. S. M., Köhler, R., Schölkopf, B., Hirsch, M.

Pattern Recognition - 38th German Conference (GCPR), 9796, pages: 426-438, Lecture Notes in Computer Science, (Editors: Rosenhahn, B. and Andres, B.), Springer International Publishing, September 2016 (conference)

ei

Arxiv Project link (url) DOI [BibTex]

Arxiv Project link (url) DOI [BibTex]


no image
Bidirektionale Interaktion zwischen Mensch und Roboter beim Bewegungslernen (BIMROB)

Kollegger, G., Ewerton, M., Peters, J., Wiemeyer, J.

11. Symposium der DVS Sportinformatik, September 2016 (conference)

ei

link (url) [BibTex]

link (url) [BibTex]


no image
A Low-cost Sensor Glove with Vibrotactile Feedback and Multiple Finger Joint and Hand Motion Sensing for Human-Robot Interaction

Weber, P., Rueckert, E., Calandra, R., Peters, J., Beckerle, P.

25th IEEE International Symposium on Robot and Human Interactive Communication (RO-MAN), pages: 99-104, August 2016 (conference)

ei

DOI [BibTex]

DOI [BibTex]


no image
Experimental and causal view on information integration in autonomous agents

Geiger, P., Hofmann, K., Schölkopf, B.

Proceedings of the 6th International Workshop on Combinations of Intelligent Methods and Applications (CIMA), pages: 21-28, (Editors: Hatzilygeroudis, I. and Palade, V.), August 2016 (conference)

ei

link (url) [BibTex]

link (url) [BibTex]


Non-parametric Models for Structured Data and Applications to Human Bodies and Natural Scenes
Non-parametric Models for Structured Data and Applications to Human Bodies and Natural Scenes

Lehrmann, A.

ETH Zurich, July 2016 (phdthesis)

Abstract
The purpose of this thesis is the study of non-parametric models for structured data and their fields of application in computer vision. We aim at the development of context-sensitive architectures which are both expressive and efficient. Our focus is on directed graphical models, in particular Bayesian networks, where we combine the flexibility of non-parametric local distributions with the efficiency of a global topology with bounded treewidth. A bound on the treewidth is obtained by either constraining the maximum indegree of the underlying graph structure or by introducing determinism. The non-parametric distributions in the nodes of the graph are given by decision trees or kernel density estimators. The information flow implied by specific network topologies, especially the resultant (conditional) independencies, allows for a natural integration and control of contextual information. We distinguish between three different types of context: static, dynamic, and semantic. In four different approaches we propose models which exhibit varying combinations of these contextual properties and allow modeling of structured data in space, time, and hierarchies derived thereof. The generative character of the presented models enables a direct synthesis of plausible hypotheses. Extensive experiments validate the developed models in two application scenarios which are of particular interest in computer vision: human bodies and natural scenes. In the practical sections of this work we discuss both areas from different angles and show applications of our models to human pose, motion, and segmentation as well as object categorization and localization. Here, we benefit from the availability of modern datasets of unprecedented size and diversity. Comparisons to traditional approaches and state-of-the-art research on the basis of well-established evaluation criteria allows the objective assessment of our contributions.

ps

pdf [BibTex]