Accelerometer-Based Gait Recognition by Sparse Representation of Signature Points With Clusters

Abstract

Gait, as a promising biometric for recognizing human identities, can be nonintrusively captured as a series of acceleration signals using wearable or portable smart devices. It can be used for access control. Most existing methods on accelerometer-based gait recognition require explicit step-cycle detection, suffering from cycle detection failures and intercycle phase misalignment. We propose a novel algorithm that avoids both the above two problems. It makes use of a type of salient points termed signature points (SPs), and has three components: 1) a multiscale SP extraction method, including the localization and SP descriptors; 2) a sparse representation scheme for encoding newly emerged SPs with known ones in terms of their descriptors, where the phase propinquity of the SPs in a cluster is leveraged to ensure the physical meaningfulness of the codes; and 3) a classifier for the sparse-code collections associated with the SPs of a series. Experimental results on our publicly available dataset of 175 subjects showed that our algorithm outperformed existing methods, even if the step cycles were perfectly detected for them. When the accelerometers at five different body locations were used together, it achieved the rank-1 accuracy of 95.8% for identification, and the equal error rate of 2.2% for verification.

Inertial Sensor-Based Gait Recognition: A Review

“…Additionally, stand-alone sensors allow for experimenting with arbitrary sensor installation which can be applied in order to examine the influence of sensor-induced factors on recognition ( i.e. , position, orientation) [ 27 , 28 , 29 ]. Furthermore, such configuration is also applicable as a supplementary part of body area network [ 30 ].…”

“…In the very first investigations, researchers experimented with relatively high sampling rates around 250 Hz [ 23 , 24 , 40 , 41 ]. In the following years, majority of stand-alone sensor-based approaches used the sampling rates in the range between 50 and 100 Hz [ 27 , 29 , 42 , 43 ]. Similarly, smartphone-based approaches relied on sampling rates below 100 Hz with most efficient approaches even using relatively low sample rate of 25 Hz [ 44 , 45 , 46 ].…”

Gait Recognition via Deep Learning of the Center-of-Pressure Trajectory

“…These sensors are used extensively in biomedical applications for gait analysis, and, hence, benefit research efforts substantially [9]. High recognition rates (>95%) have been observed under laboratory conditions [10], but a lower accuracy (~70%) has been reported in more realistic datasets [11].…”

Pattern Mining Approaches Used in Sensor-Based Biometric Recognition: A Review

“…A grid-based structure is constructed by quantizing the biometric feature space into limited number of cells [135], i.e. 1) the data space is initially divided into certain number of cells, 2) the cell density for every of the cell is computed, 3) cells are categorized through sorting according to their densities, 4) the center of the cluster is acknowledged, 5) the distance amongst the neighboring cells are computed.…”