Unifying Representations and Large-Scale Whole-Body Motion Databases for Studying Human Motion

Mandery, Christian; Terlemez, Ömer; Do, Martin; Vahrenkamp, Nikolaus; Asfour, Tamim

doi:10.1109/tro.2016.2572685

Cited by 90 publications

(53 citation statements)

References 38 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The most common methods to estimate human intention are mainly based on minimum jerk models [2], or on imitation learning techniques. In the latter, the movements of two human actors are typically retrieved with motion capture techniques, clustered in motion databases [3], [4], [5] and then used to learn the interaction skills [6], [7], [8]. In contrast to these approaches, we formulate the problem as the following: given a human model, and measurements provided by wearable sensors, we want to estimate the current human configuration (state) and intention (dynamical quantities: force-torques, accelerations, etc.…”

Section: References Controllermentioning

confidence: 99%

The CoDyCo Project Achievements and Beyond: Toward Human Aware Whole-Body Controllers for Physical Human Robot Interaction

Romanò

Nava

Azad

et al. 2018

IEEE Robot. Autom. Lett.

View full text Add to dashboard Cite

The success of robots in real-world environments is largely dependent on their ability to interact with both humans and said environment. The FP7 EU project CoDyCo focused on the latter of these two challenges by exploiting both rigid and compliant contacts dynamics in the robot control problem. Regarding the former, to properly manage interaction dynamics on the robot control side, an estimation of the human behaviours and intentions is necessary. In this paper we present the building blocks of such a human-in-the-loop controller, and validate them in both simulation and on the iCub humanoid robot using a human-robot interaction scenario. In this scenario, a human assists the robot in standing up from being seated on a bench. Index Terms-Physical Human-Robot Interaction, Humanoid Robots I. INTRODUCTION T HE ability to interact with and manipulate the environment gives robots a distinct advantage over purely software based automated agents. In the FP7 European project, CoDyCo, the focus was on how to properly exploit contact dynamics in the control of the robot. When the interaction involves humans, their intrinsic unpredictability makes the collaboration problem far more difficult. Foreseen robotic applications range from the use of robots as service and elderly assistants, to their use in industrial plants in close contact with

show abstract

Section: References Controllermentioning

confidence: 99%

The CoDyCo Project Achievements and Beyond: Toward Human Aware Whole-Body Controllers for Physical Human Robot Interaction

Romanò

Nava

Azad

et al. 2018

IEEE Robot. Autom. Lett.

View full text Add to dashboard Cite

show abstract

“…The Human3.6M Dataset [11] provides a large-scale dataset for the evaluation for human pose recognition methods, which contains complementary data from time-of-flight cameras and 3D laser scans of the human subjects. Mandery et al [5] provide a more indepth discussion of existing human motion databases regarding size, methodology, and available motion types, also including databases that have specialized on more specific types of motion instead of whole-body actions.…”

Section: Related Workmentioning

confidence: 99%

“…For this purpose, we created a web-based tool specifically designed for this task called the Motion Annotation Tool. 5 We describe the user interface of the tool, cover gamification approaches that we used to motivate annotators, and finally describe a novel approach that we used to decide which motion to present next to the user for annotation. The source code of the Motion Annotation Tool is openly available.…”

Section: Natural Language Annotationsmentioning

confidence: 99%

The KIT Motion-Language Dataset

2016

Self Cite

View full text Add to dashboard Cite

Linking human motion and natural language is of great interest for the generation of semantic representations of human activities as well as for the generation of robot activities based on natural language input. However, although there have been years of research in this area, no standardized and openly available data set exists to support the development and evaluation of such systems. We, therefore, propose the Karlsruhe Institute of Technology (KIT) Motion-Language Dataset, which is large, open, and extensible. We aggregate data from multiple motion capture databases and include them in our data set using a unified representation that is independent of the capture system or marker set, making it easy to work with the data regardless of its origin. To obtain motion annotations in natural language, we apply a crowd-sourcing approach and a web-based tool that was specifically build for this purpose, the Motion Annotation Tool. We thoroughly document the annotation process itself and discuss gamification methods that we used to keep annotators motivated. We further propose a novel method, perplexity-based selection, which systematically selects motions for further annotation that are either under-represented in our data set or that have erroneous annotations. We show that our method mitigates the two aforementioned problems and ensures a systematic annotation process. We provide an in-depth analysis of the structure and contents of our resulting data set, which, as of October 10, 2016, contains 3911 motions with a total duration of 11.23 hours and 6278 annotations in natural language that contain 52,903 words. We believe this makes our data set an excellent choice that enables more transparent and comparable research in this important area.

show abstract

“…To gather data, the KIT Whole-Body Human Motion Database, introduced in [31] and [32], was browsed for suitable motion recordings. The criteria were that the recordings are relatively long and feature at least two different activities.…”

Section: B Frameworkmentioning

confidence: 99%

A framework for evaluating motion segmentation algorithms

Dreher

Kulp

Mandery

et al. 2017

2017 IEEE-RAS 17th International Conference on Humanoid Robotics (Humanoids)

Self Cite

View full text Add to dashboard Cite

There have been many proposals for algorithms segmenting human whole-body motion in the literature. However, the wide range of use cases, datasets, and quality measures that were used for the evaluation render the comparison of algorithms challenging. In this paper, we introduce a framework that puts motion segmentation algorithms on a unified testing ground and provides a possibility to allow comparing them. The testing ground features both a set of quality measures known from the literature and a novel approach tailored to the evaluation of motion segmentation algorithms, termed Integrated Kernel approach. Datasets of motion recordings, provided with a ground truth, are included as well. They are labelled in a new way, which hierarchically organises the ground truth, to cover different use cases that segmentation algorithms can possess. The framework and datasets are publicly available and are intended to represent a service for the community regarding the comparison and evaluation of existing and new motion segmentation algorithms.

show abstract

Unifying Representations and Large-Scale Whole-Body Motion Databases for Studying Human Motion

Cited by 90 publications

References 38 publications

The CoDyCo Project Achievements and Beyond: Toward Human Aware Whole-Body Controllers for Physical Human Robot Interaction

The CoDyCo Project Achievements and Beyond: Toward Human Aware Whole-Body Controllers for Physical Human Robot Interaction

The KIT Motion-Language Dataset

A framework for evaluating motion segmentation algorithms

Contact Info

Product

Resources

About