Font Size: a A A

Based On Gesture Recognition By Used Mobile Phone Universal Control

Posted on:2011-06-11Degree:MasterType:Thesis
Country:ChinaCandidate:Y DingFull Text:PDF
GTID:2178330332964071Subject:Computer application technology
Abstract/Summary:PDF Full Text Request
In the information age, the "pervasive computing" has been widely used, through pervasive computing, which can provide equipment utilization, so that people can get more convenient service and more effective in life. HCI is one of the most hot topics the study of ubiquitous computing, and natural gesture recognition is one of the hotspots in human-computer interaction research. Our studies main purpose is to bring the results of gesture recognition applied to human-computer interaction interface, thus achieving a more natural and harmonious human-computer interaction. For instance, we can let gesture recognition apply in intelligent control, scene services, virtual reality, 3D design, visualization, medical research and other fields of study. Through the application of these research results we can be achieved good social and economic benefits.This study is based on "ubiquitous computing" environment, to achieve common control on novel type of media control interface which based on natural gestures recognition by signaled cell phone and data transmission services based on Bluetooth. User express their operation intended by waving cell phone which is in-built three-axis acceleration sensor that can obtain the data of appropriate gesture and the data of gesture mode can transmit to the server by the Bluetooth on real-time. The server obtains the data model of gesture and distinguish on real-time. The gesture recognition results will be match common control commands, and finally achieve universal control system is simple and convenient operation. This article includes the following areas:First: This paper fully exploit and utilize a common cell phone features, and select the phone as an interaction equipment of system. Compared with other equipment is more convenient, practical and reflects the benefits of ubiquitous computing generally applicable. Designed and implemented the acceleration data of collection and Bluetooth data transfer software based on mobile.Second: Through reading the literature of gesture recognition at home and abroad, select the DTW, HMM and SVM three classic gesture recognition algorithms to make the experimental comparison on training model. Analyzing three algorithms from the number of training samples needs, training of pretreatment, and training time complexity, select the most simples DTW algorithm as a system identification algorithm, then this algorithm improvement for the DTW-MT algorithm. Experiments show: Using DTW-MT algorithm is best for the system identification algorithm, based on 5 template DTW-MT algorithm the time required 14.4ms to identify a sample, the identification algorithm can reduce the operation frequency which user change the template in the actual, raising the average rate of recognition, on the phone several common gesture of the average recognition rate can reach 98.2%. This algorithm use in general control systems can achieve real-time identification and common control.Finally: This paper implements two general-purpose control systems through the mobile phone acceleration and Bluetooth module combined with the DTW-MT gesture recognition algorithm: (1) phone-based gesture recognition of new media control interface, the system achieve a simple, convenient, practical common control on the new media machine for user. Users can achieve the media machine of the player directory switching, fast forward, increase or decrease the sound, etc. by simply waving the phone; (2) Bluetooth data transmission service based on mobile phone common control system, the system to implement user access server file by using a mobile phone, and access to the server files or simple file operations by means of gesture recognition, such as it can be achieved the server file transfer, copy and delete operations by waving mobile phones.The study implements two general-purpose control systems by phone features combined with gesture recognition algorithm, which can provide users with simple, easy operation of common control.
Keywords/Search Tags:Dynamic time warping (DTW), Hidden Markov Model (HMM), Support Vector Machine (SVM), Multi-Template dynamic time warping (DTW-MT), natural gesture recognition, human-computer interaction, media player interface
PDF Full Text Request
Related items