Using Hidden Markov Models for accelerometer-based biometric gait recognition

Abstract

Abstract-Biometric gait recognition based on accelerometer data is still a new field of research. It has the merit of offering an unobtrusive and hence user-friendly method for authentication on mobile phones. Most publications in this area are based on extracting cycles (two steps) from the gait data which are later used as features in the authentication process. In this paper the application of Hidden Markov Models is proposed instead. These have already been successfully implemented in speaker recognition systems. The advantage is that no error-prone cycle extraction has to be performed, but the accelerometer data can be directly used to construct the model and thus form the basis for successful recognition. Testing this method with accelerometer data of 48 subjects recorded using a commercial of the shelve mobile phone a false non match rate (FNMR) of 10.42% at a false match rate (FMR) of 10.29% was obtained. This is half of the error rate obtained when applying an advanced cycle extraction method to the same data set in previous work.

An Overview of Smart Shoes in the Internet of Health Things: Gait and Mobility Assessment in Health Promotion and Disease Monitoring

“…The respective approaches use algorithms such as support vector machines [58], artificial neural networks [59], hidden Markov model classifiers [60], and Gaussian mixture model classifiers [61]. A detailed report on these and further classifiers can be found in a review paper on activity classification [62].…”

Analysis of Image Classification and Annotation Methods for Smartphone

“…Year Ref Classification Algorithms Evaluation Method Processing Platform 2014 [1] Multidimensional sensing heuristic correlate to accelerometer measurements Phone 2014 [2] MPEG compact descriptors for visual search Phone 2011 [19] Hidden Markov Models 4-fold cross validation and precision/recall measures Phone 2011 [17] Multiclass Logistic Regression 4-fold cross validation and precision/recall measures Server 2011 [20] Transfer learning Embedded Decision Tree 10 times 10-fold cross validation Server 2011 [28] Hidden Markov Chain Server 2011 [18] Decision Tree, Naïve Bayes, Random Forest, Logistics Regression, RBF Network, Support Vector Machine 10-fold cross validation Server 2011 [30] Smoothed Single-layer Hidden Markov Models F-measure Server 2011 [14] Decision Tree Phone 2011 [16] Hidden Markov Model false non match rate (FNMR), false match rate (FMR)…”

Gait Identification using Smartphone Handheld Placement with Linear Interpolation Factor, Single Magnitude and One-vs-One Classifier Mapping

“…To overcome this situation, the obtained data need to be increased its sampling rate depending on the number of samples per second (Hz) which also called as interpolation. According to [11], there are generally 2 methods in implementing interpolation on the gait signal which is linear [12][13][14][15][16][17][18] and cubic spline [12]. There are also papers that do not mentioned the use of any interpolation in gait application [19][20].…”