Real time gesture recognition using continuous time recurrent neural networks

Bailador, Gonzalo; Roggen, Daniel; Tröster, Gerhard; Triviño, Gracián

doi:10.4108/bodynets.2007.149

Cited by 58 publications

(29 citation statements)

References 23 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…However, the use of Recurrent Neural Networks has been restricted to human action recognition studies (Bailador, Roggen, Tröester & Triviño 2007;Baccouche, Mamalet, Wolf, Garcia, & Baskurt, 2011;Lee & Cho, 2012;Tan & De Silva, 2003, among others). Thus, the objective of this paper is to compare the performance of different classifiers to recognize transportation modes, incorporating Recurrent Neural Networks among them.…”

Section: Processing Of the Data And Identification Of Transportationmentioning

confidence: 98%

Travel Behavior Characterization Using Raw Accelerometer Data Collected from Smartphones

Ferrer

Ruíz

2014

Procedia - Social and Behavioral Sciences

View full text Add to dashboard Cite

In this paper, we compare different algorithms for the recognition of transportation modes based on features extracted from the accelerometer data. The performance and effectiveness of the transportation mode classifiers presented is evaluated and their accuracy is discussed. The data set used for training and testing algorithms was collected by a group of volunteers in the city of Valencia in 2013; an Android application designed for the recording of trips and transportation modes application was installed on their smartphones. This application collected GPS readings each 10-12 seconds and accelerometer data at 1Hz. While GPS data was only used for the validation of trips for the training of the algorithms, accelerometer readings were used entirely for their training. Results show the high performance of Recurrent Neural Networks in recognizing travel modes using accelerometer data.

show abstract

Section: Processing Of the Data And Identification Of Transportationmentioning

confidence: 98%

Travel Behavior Characterization Using Raw Accelerometer Data Collected from Smartphones

Ferrer

Ruíz

2014

Procedia - Social and Behavioral Sciences

View full text Add to dashboard Cite

show abstract

“…A recognition rate of 90.45% was achieved, using a set of 31 different gestures, with a mean processing time of 1.5 seconds per gesture. Bailador et al [16] recognized dynamic gestures using a set of Continuous Time Recurrent Neural Networks (CTRNN). The gesture data were generated by accelerometers, and the neural networks were used to predict future acceleration values from current ones; the gesture related to the neural network with the smallest prediction error wins.…”

Section: Neural Networkmentioning

confidence: 99%

Comparative evaluation of static gesture recognition techniques based on nearest neighbor, neural networks and support vector machines

Savaris

Wangenheim

2010

J Braz Comput Soc

View full text Add to dashboard Cite

It is a common behavior for human beings to use gestures as a means of expression, as a complement to speaking, or as a self-contained communication mode. In the field of Human-Computer Interaction, this behavior can be adopted to build alternative interfaces, aiming to ease the relationship between the human element and the computational element. Currently, various gesture recognition techniques are described in the technical literature; however, the validation studies of these techniques are usually performed isolatedly, which complicates comparisons between them. To reduce this gap, this work presents a comparison between three well-established techniques for static gesture recognition, using Nearest Neighbor, Neural Networks, and Support Vector Machines as classifiers. These classifiers evaluate a common dataset, acquired from an instrumented glove, and generate results for precision and performance measurements. The results obtained show that the classifier implemented as a Support Vector Machine presented the best generalization, with the highest recognition rate. In terms of performance, all methods presented evaluation times fast enough to be used interactively. Finally, this work identifies and discusses a set of relevant criteria that must be observed for the training and evaluation steps, and its relation to the final results.

show abstract

“…The prediction-errorclassification approach [31], [32] has been proposed which builds predictive models based either on neuro-fuzzy predictors or continuous-time recurrent neural networks. These models learn to predict the acceleration values in the X, Y and Z axes in the next time step for eight different gestures.…”

Section: Prediction Ahead In Timementioning

confidence: 99%

Prediction-Based Audiovisual Fusion for Classification of Non-Linguistic Vocalisations

Petridis

Pantić

2016

IEEE Trans. Affective Comput.

View full text Add to dashboard Cite

Abstract-Prediction plays a key role in recent computational models of the brain and it has been suggested that the brain constantly makes multisensory spatiotemporal predictions. Inspired by these findings we tackle the problem of audiovisual fusion from a new perspective based on prediction. We train predictive models which model the spatiotemporal relationship between audio and visual features by learning the audio-to-visual and visual-to-audio feature mapping for each class. Similarly, we train predictive models which model the time evolution of audio and visual features by learning the past-to-future feature mapping for each class. In classification, all the class-specific regression models produce a prediction of the expected audio / visual features and their prediction errors are combined for each class. The set of class-specific regressors which best describes the audiovisual feature relationship, i.e., results in the lowest prediction error, is chosen to label the input frame. We perform cross-database experiments, using the AMI, SAL and MAHNOB databases, in order to classify laughter and speech and subject-independent experiments on the AVIC database in order to classify laughter, hesitation and consent. In virtually all cases prediction-based audiovisual fusion consistently outperforms the two most commonly used fusion approaches, decision-level and feature-level fusion.

show abstract

Real time gesture recognition using continuous time recurrent neural networks

Cited by 58 publications

References 23 publications

Travel Behavior Characterization Using Raw Accelerometer Data Collected from Smartphones

Travel Behavior Characterization Using Raw Accelerometer Data Collected from Smartphones

Comparative evaluation of static gesture recognition techniques based on nearest neighbor, neural networks and support vector machines

Prediction-Based Audiovisual Fusion for Classification of Non-Linguistic Vocalisations

Contact Info

Product

Resources

About