A comparative study of classification methods for gesture recognition using a 3-axis accelerometer

Abstract

We used Fisher linear discriminant analysis (LDA), static neural networks (NN), and focused time delay neural networks (TDNN) for gesture recognition. Gestures were collected in form of acceleration signals along three axes from six participants. A sports watch containing a 3-axis accelerometer, was worn by the users, who performed four gestures. Each gesture was performed for ten seconds, at the speed of one gesture per second. User-dependent and userindependent k-fold cross validations were carried out to measure the classifier performance. Using first and second order statistical descriptors of acceleration signals from validation datasets, LDA and NN classifiers were able to recognize the gestures at an average rate of 86% and 97% (user-dependent) and 89% and 85% (user-independent), respectively. TDNNs proved to be the best, achieving near perfect classification rates both for user-dependent and userindependent scenarios, while operating directly on the acceleration signals alleviating the need for explicit feature extraction.

Human activity recognition using multisensor data fusion based on Reservoir Computing

“…Several works in literature report examples of applications of delay neural networks to problems in the field of human activity recognition, taking into consideration temporal windows with length typically in the range of 2-10 seconds. For example, in [55,56] delay networks are used to classify daily-life activities from accelerometer data, whereas a similar approach is used in [57] for human gesture recognition, and in [58] for recognition of workers activities. In [59] a delay network based approach is used to recognize human postures and activities from data collected by a smart-shoe device.…”

Natural hand gestures for human identification in a Human-Computer Interface

“…Authors show that such gestures can be effectively presented in the bi-dimensional space. In [26], the LDA classifier was compared with neural networks (NN) and focused time delay neural networks (TDNN) for gesture recognition based on data from a 3-axis accelerometer. LDA gave similar results to the NN approach, and the TDNN technique, though computationally more complex, achieved better performance.…”

Accelerometer Based Static Gesture Recognition and Mobile Monitoring System Using Neural Networks

“…: +91-8148736577; fax: +91-44 3993 2555e-mail address: maddikerakalyan@vit.ac.in Accelerometer based industrial robotic arms are already in use and the studies in this view have been in prevalence 5 .The gestures of different organs of the body are used to control the wheel chair and different intelligent mechanisms have been developed to control the intermediate mechanisms 6 .Hidden markov models and neural networks have always imparted a sense of intelligence to the controlling and classifying mechanisms in different industrial applications 7 .accelorometers have also involved themselves in to the digital and hand written character recognition based on gesture classification 8 .The application of 3-axis accelerometers signal classification involves quite a very precision and uses different methods. The comparison among these classification methods has always been debatable 9 .LVQ is one of those mostly used methods, which proves to be given robust classification among many complicated patterns 10 . Many neural network algorithms have been repetitively applied for the purpose of classifying the signatures and gestures.…”