Feature design scheme for Kinect-based DTW human gesture recognition

Ding, Ing-Jr; Chang, Che-Wei

doi:10.1007/s11042-015-2782-3

Cited by 22 publications

(17 citation statements)

References 11 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Human gesture recognition is mainly performed on RGB-D (Red, Green, Blue and Depth) data [9], [10], [11], [12], [6], [13] or on skeleton data [14], [15], [16], [17], [18], [13], [19], [20], [21], where skeletal data can be extracted from RGB-D data. To recognize static gestures (i.e.…”

Section: Movementioning

confidence: 99%

Human gesture classification by brute-force machine learning for exergaming in physiotherapy

Deboeverie

Roegiers

Allebosch

et al. 2016

2016 IEEE Conference on Computational Intelligence and Games (CIG)

View full text Add to dashboard Cite

Abstract-In this paper, a novel approach for human gesture classification on skeletal data is proposed for the application of exergaming in physiotherapy. Unlike existing methods, we propose to use a general classifier like Random Forests to recognize dynamic gestures. The temporal dimension is handled afterwards by majority voting in a sliding window over the consecutive predictions of the classifier. The gestures can have partially similar postures, such that the classifier will decide on the dissimilar postures. This brute-force classification strategy is permitted, because dynamic human gestures show sufficient dissimilar postures. Online continuous human gesture recognition can classify dynamic gestures in an early stage, which is a crucial advantage when controlling a game by automatic gesture recognition. Also, ground truth can be easily obtained, since all postures in a gesture get the same label, without any discretization into consecutive postures. This way, new gestures can be easily added, which is advantageous in adaptive game development. We evaluate our strategy by a leave-one-subject-out cross-validation on a self-captured stealth game gesture dataset and the publicly available Microsoft Research Cambridge-12 Kinect (MSRC-12) dataset. On the first dataset we achieve an excellent accuracy rate of 96.72%. Furthermore, we show that Random Forests perform better than Support Vector Machines. On the second dataset we achieve an accuracy rate of 98.37%, which is on average 3.57% better then existing methods.

show abstract

Section: Movementioning

confidence: 99%

Human gesture classification by brute-force machine learning for exergaming in physiotherapy

Deboeverie

Roegiers

Allebosch

et al. 2016

2016 IEEE Conference on Computational Intelligence and Games (CIG)

View full text Add to dashboard Cite

show abstract

Section: Summary Of 2accepted Papersmentioning

confidence: 99%

“…The first seven papers belong to the technique category of smart pattern recognition [4,7,10,13,14,18,22]. The first paper entitled BFeature design scheme for Kinect-based DTW human gesture recognition^by Ding and Chang [7] develops a feature design scheme involving hybridizations of joint positions and joint angles for dynamic time warping-based human gesture recognition with the Kinect camera. The second paper entitled BImplementation of an interactive TV interface via gesture and handwritten numeral recognition^by Sheu and Huang [14] considers the utilization of the Kinect sensor to recognize the dynamic gestures of users to achieve real-time interactive control of TV where a back-propagation neural network was also used to perform handwriting recognition to help the command recognition task.…”

Section: Summary Of 2accepted Papersmentioning

confidence: 99%

Guest Editorial: Smart Multimedia Communication Technologies for New-Generation Artificial Intelligence Applications

Hsieh

2016

Multimed Tools Appl

View full text Add to dashboard Cite

With the development of the Internet of thing (IoT) mechanism nowadays, the required techniques for artificial intelligence (AI)-related applications will be equipped with a crucial property, which is Bsmart.^Smart multimedia communication technologies are composed of two main technique categories: the smart multimedia processing technique class and the smart communication transmission technique class. These will be the necessity for constructing a new-generation AI application. The technical category of multimedia processing involves lots of types of techniques where pattern recognition, video (or image) processing, and human machine interaction (HCI) are three kernel technique types to establish an AI application under the regulation of the IoT mechanism. Video coding and decoding (codecs) techniques, media streaming techniques, and hardware-based techniques for enhancing media communication play important roles for establishing an efficient communication transmission channel between the user end device and the remote media handling server.BSmart multimedia communication technologies^associated with the new-generation AI application will absolutely be an evolution tendency for the development of multimedia and communication technologies. Following this thought line, we deeply believe that the appropriate term Bsmart^will be given to pattern recognition, video processing, human machine interaction, media streaming, and hardware-based media communication enhancement techniques for building up a smart multimedia communication system.The special issue, BSmart Multimedia Communication Technologies for New-Generation Artificial Intelligence Applications,^acts as a path breaker for exploring new-generation AI applications using multimedia communication technologies with intelligence. We totally received around 30 submissions, and the vast majority of all submissions are significant in Multimed Tools Appl (2016) 75:9665-9668

show abstract

“…The Kinect sensor and its software development kit (Kinect SDK) can be efficiently used to simplify the conventional gesture recognition technique. Recently, studies on Kinect-based gesture recognition have been frequently seen [3][4][5][6], and Kinect-based gesture recognition applications mainly include robot automation, health-care, entertainment and surveillance. For the vending machine operation application using Kinect-based gesture recognition, extremely rare studies are done.…”

Section: *Corresponding Author: Ing-jr Ding Department Of Electrical mentioning

confidence: 99%

“…A hidden Markov model (HMM) approach for Kinect-based gesture recognition is proposed in [4] where the recognition accuracy of the HMM gesture model could be improved by properly adjusting HMM model parameters. In addition, in [5], a feature design scheme is developed for the computation of Kinect-based dynamic time warping gesture recognition. All those gesture recognition computations without the utilization of the Kinect software development kit in [3][4][5] will be able to improve the recognition performance of Kinect-based gesture recognition by the self-developed gesture recognition method, the gesture model adjustment scheme, or the accurate gesture feature for recognition.…”

Section: A Gesture Recognition Computations With/without Kinect Sdkmentioning

confidence: 99%

Kinect based Gesture Command Recognition for Vending Machine Operations Using an Android Platform as the Machine Simulator

Ding¹,

Lin²,

Wu³

2015

Second International Conference on Advances in Computing, Control and Networking - ACCN 2015

View full text Add to dashboard Cite

In this paper, a Kinect-based gesture recognition scheme for the application of operating a vending machine is presented where the popular smart phone device with the android platform is used as a simulator of the vending machine. Compared to the conventional human-computer interface (HCI) of vending machines by pushbutton operations, the developed gesture recognition scheme that employs the gesture action made by a personal operator as the operational command is a new type of HCI to control the vending machine. Presented Kinect-based gesture recognition for vending machine control sufficiently takes use of the released Kinect software development kit (Kinect SDK) to develop the system where the human skeleton information to represent the corresponding operator's gesture action is extracted and then recognized. For further increasing the recognition accuracy of gesture command recognition, a time interval between the current recognition decision and the next recognition decision is investigated and analyzed. A proper value of decision time-interval will then be provided to the gesture recognition system for avoiding the imperfect recognition accuracy that results from an arbitrary and improper setting of the decision time-interval. Experimental results showed that presented Kinect-based gesture command recognition for vending machine operations is feasible and also competitive on recognition performances. On the use of the released Kinect SDK for establishing the gesture recognition system, gesture recognition with an appropriately-designed time interval of decision making will perform apparently much better than that without any investigation information available about decision time-interval settings.

show abstract

Feature design scheme for Kinect-based DTW human gesture recognition

Cited by 22 publications

References 11 publications

Human gesture classification by brute-force machine learning for exergaming in physiotherapy

Human gesture classification by brute-force machine learning for exergaming in physiotherapy

Guest Editorial: Smart Multimedia Communication Technologies for New-Generation Artificial Intelligence Applications

Kinect based Gesture Command Recognition for Vending Machine Operations Using an Android Platform as the Machine Simulator

Contact Info

Product

Resources

About