A New Computational Approach to Identify Human Social Intention in Action

Daoudi, Mohamed; Yann, C.; Descrosiers, Paul-Audain; Ott, Laurent

doi:10.1109/fg.2018.00082

Cited by 9 publications

(8 citation statements)

References 12 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…To achieve this, we used the computational method developed by Daoudi et al (2018) , which allowed us to analyse online participants' grasp-toplace actions and, on the basis of their kinematic pro le, to categorise them as being driven by either a social or a personal intention. The reaction of the virtual agent would therefore serve as feedback allowing participants to progressively adapt their motor performances and succeed the task.…”

Section: Experiments 2 Inducing the Effects Of Social Intention In Hu...mentioning

confidence: 99%

“…The task, the experimental setup and the stimuli were the same as used in Experiment 1, except for the reaction of the virtual agent, that was not set a priori but depended on the kinematic features of the participants' motor actions. For this purpose, a classi cation algorithm (Daoudi et al, 2018;Devanne et al, 2015) was used to analyse participants' trajectories online and classify them as driven by a social or a personal intention on the basis of their speci c kinematic pro le. Once the intention was classi ed by the algorithm, the virtual agent could either ll the glass of the participants (if the movement was classi ed as driven by a social intention) or not (if the movement was classi ed as driven by a personal intention).…”

Section: Materials and Stimulimentioning

confidence: 99%

“…The present section describes brie y the algorithm used to analyse participants' actions and classify them as driven by either a personal or a social intention (for more details, see Daoudi et al, 2018;Devanne et al, 2015) . The 3D trajectories of each action were rst represented as curves in ℝ 3 by means of the Square-Root Velocity Function (SRVF; Srivastava et al, 2010) .…”

Section: Description Of the Classi Cation Algorithmmentioning

confidence: 99%

See 2 more Smart Citations

Different effects of social intention on movement kinematics when interacting with a human or virtual agent

Gigliotti,

Desrosiers,

Ott

et al. 2023

Preprint

Self Cite

View full text Add to dashboard Cite

The execution of object-directed motor actions is known to be influenced by the intention to interact with others. In this study, we tested whether the effects of social intention on the kinematics of object-directed actions depended on whether the task was performed in the presence of a human or a virtual confederate. In two experiments, participants had to grasp a glass and place it to a new position, with either a personal intention (to fill the glass themselves using a bottle) or a social one (to have the glass filled by the human confederate or the virtual agent using the bottle). Experiment 1 showed that the kinematics of the object-directed actions was modulated by the social intention but only when interacting with a human confederate. Experiment 2 showed that the effects of social intention on object-directed actions performance can be improved using feedback-based learning. Specifically, participants proved able to progressively adapt their motor performances to express social intention to a virtual confederate as well. These findings emphasize the importance of the modulatory role of social intention on non-verbal motor behaviour, and enrich the understanding of the interaction with virtual agents.

show abstract

Section: Experiments 2 Inducing the Effects Of Social Intention In Hu...mentioning

confidence: 99%

Section: Materials and Stimulimentioning

confidence: 99%

Section: Description Of the Classi Cation Algorithmmentioning

confidence: 99%

See 1 more Smart Citation

Different effects of social intention on movement kinematics when interacting with a human or virtual agent

Gigliotti,

Desrosiers,

Ott

et al. 2023

Preprint

Self Cite

View full text Add to dashboard Cite

show abstract

“…However, the majority of the research concentrated on classifying human actions rather than predicting the numerous human motivations that drive human behavior. Many more uses of intention prediction have been considered, including driving systems, rehabilitation, and security (Daoudi et al, 2018). The distinction of this work is to predict intention from vision only, using RGB images of human activity.…”

Section: Literature Reviewmentioning

confidence: 99%

Deep Learning-based human activity recognition using RGB images in Human-robot collaboration

Keshinro

Seong

Sun

2022

Proceedings of the Human Factors and Ergonomics Society Annual

View full text Add to dashboard Cite

In human-robot interaction, to ensure safety and effectiveness, robots need to be able to accurately predict human intentions. Hidden Markov Model, Bayesian Filtering, and deep learning methods have been used to predict human intentions. However, few studies have explored deep learning methods to predict variant human intention. Our study aims to evaluate the performance of the human intent recognition inference algorithm, and its impact on the human-robot team for collaborative tasks. Two deep learning algorithms ConvLSTM and LRCN were used to predict human intention. A dataset of 10 participants performing Pick, Throw, Wave, and Carry actions was used. The ConvLSTM method had a prediction accuracy of 74%. The LRCN method had a lower prediction accuracy of 25% compared to ConvLSTM. This result shows that deep learning methods using RGB images can predict human intent with high accuracy. The proposed method is successful in predicting human intents underlying human behavior.

show abstract

“…POMDP [1] was used to select anticipatory actions in human-robot table tennis playing. By analyzing the trajectories of human arms, the social intention prediction problem was investigated in [27]. Intention prediction was also studied in [10] using motion information (i.e., kinematic data and video data) only.…”

Section: Related Workmentioning

confidence: 99%

Untitled

2022

IJAIA

View full text Add to dashboard Cite

Data augmentation has been broadly applied in training deep-learning models to increase the diversity of data. This study ingestigates the effectiveness of different data augmentation methods for deep-learningbased human intention prediction when only limited training data is available.A human participant pitches a ball to nine potential targets in our experiment. We expect to predict which target the participant pitches the ball to. Firstly, the effectiveness of 10 data augmentation groups is evaluated on a single-participant data set using RGB images. Secondly, the best data augmentation method (i.e., random cropping) on the single-participant data set is further evaluated on a multi-participant data set to assess its generalization ability. Finally, the effectiveness of random cropping on fusion data of RGB images and optical flow is evaluated on both single-and multi-participant data sets. Experiment results show that: 1) Data augmentation methods that crop or deform images can improve the prediction performance; 2) Random cropping can be generalized to the multi-participant data set (prediction accuracy is improved from 50% to 57.4%); and 3) Random cropping with fusion data of RGB images and optical flow can further improve the prediction accuracy from 57.4% to 63.9% on the multi-participant data set.

show abstract

A New Computational Approach to Identify Human Social Intention in Action

Cited by 9 publications

References 12 publications

Different effects of social intention on movement kinematics when interacting with a human or virtual agent

Different effects of social intention on movement kinematics when interacting with a human or virtual agent

Deep Learning-based human activity recognition using RGB images in Human-robot collaboration

Untitled

Contact Info

Product

Resources

About