Asif Mehmood

Extraction of the velocity of walking human’s body segments using ultrasonic Doppler

The focus of this paper is to experimentally extract the Doppler signatures of a walking humans individual body segments using an ultrasonic Doppler system (UDS) operating at 40 kHz. In a humans walk, the major contribution to Doppler velocities and acoustic scattering is from the foot, lower leg, thigh (upper leg) and torso. The Doppler signature of these human body segments are extracted experimentally. The measurements were made by illuminating one of these body segments at a time and blocking the remaining body segments using acoustic screens. The results obtained in our experiment were verified with the results published by Bradley using a physics-based model for Doppler sonar spectrograms.

Separation of human and animal seismic signatures using non-negative matrix factorization

Seismic footstep detection based systems can be employed for homeland security applications such as perimeter protection and the border security. This paper reports an approach based on non-negative matrix factorization (NMF) for seismic footstep signal separation for a single channel recording. A supervised NMF technique is employed to separate the human footstep signatures from the horse footstep signatures. The proposed algorithm is applied on the spectrogram of human footstep signals and horse footstep signals. The spectrograms of these signals are presented as a sum of components, each having a fixed spectrum and time-varying gain. The main benefit of the proposed technique is its ability to decompose a complex signal automatically into objects that have a meaningful interpretation. In this paper, a sparsity-based NMF algorithm is developed and implemented on seismic data of human and horse footsteps. The performance of this method is very promising and is demonstrated by the experimental results.

Discrimination of bipeds from quadrupeds using seismic footstep signatures

Seismic sensors are widely used to detect moving targets in the ground sensor network, and can be easily employed to discriminate human and quadruped based on their footstep signatures. Because of the complex environmental conditions and the non-stationary nature of the seismic signals, footstep detection and classification is a very challenging problem. The solution to this problem has various applications such as border security, surveillance, perimeter protection and intruder detection. Previous works in the domain of seismic detection of human vs. quadruped have relied on the cadence frequency-based models. However, cadence-based detection alone results in high false alarms. In this paper, we describe a seismic footstep database and present classification results based on support vector machine (SVM). We demonstrate that in addition to applying a good classification algorithm, finding robust features are very important for seismic discrimination.

Wavelet-RX anomaly detection for dual-band forward-looking infrared imagery

This paper describes a new wavelet-based anomaly detection technique for a dual-band forward-looking infrared (FLIR) sensor consisting of a coregistered longwave (LW) with a midwave (MW) sensor. The proposed approach, called the wavelet-RX (Reed-Xiaoli) algorithm, consists of a combination of a two-dimensional (2D) wavelet transform and a well-known multivariate anomaly detector called the RX algorithm. In our wavelet-RX algorithm, a 2D wavelet transform is first applied to decompose the input image into uniform subbands. A subband-image cube is formed by concatenating together a number of significant subbands (high-energy subbands). The RX algorithm is then applied to the subband-image cube obtained from a wavelet decomposition of the LW or MW sensor data. In the case of the dual band, the RX algorithm is applied to a subband-image cube constructed by concatenating together the high-energy subbands of the LW and MW subband-image cubes. Experimental results are presented for the proposed wavelet-RX and the classical constant false alarm rate (CFAR) algorithm for detecting anomalies (targets) in a single broadband FLIR (LW or MW) or in a coregistered dual-band FLIR sensor. The results show that the proposed wavelet-RX algorithm outperforms the classical CFAR detector for both single-band and dual-band FLIR sensors.

Ultrasonic Doppler methods to extract signatures of a walking human

Extraction of Doppler signatures that characterize human motion has attracted a growing interest in recent years. These Doppler signatures are generated by various components of the human body while walking, and contain unique features that can be used for human detection and recognition. Although, a significant amount of research has been done in radio frequency regime for human Doppler signature extraction, considerably less has been done in acoustics. In this work, 40 kHz ultrasonic sonar is employed to measure the Doppler signature generated by the motion of body segments using different electronic and signal processing schemes. These schemes are based on both analog and digital demodulation with homodyne and heterodyne receiver circuitry. The results and analyses from these different schemes are presented.

Kernel wavelet-Reed–Xiaoli: an anomaly detection for forward-looking infrared imagery

This paper describes a new kernel wavelet-based anomaly detection technique for long-wave (LW) forward-looking infrared imagery. The proposed approach called kernel wavelet-Reed-Xiaoli (wavelet-RX) algorithm is essentially an extension of the wavelet-RX algorithm (combination of wavelet transform and RX anomaly detector) to a high-dimensional feature space (possibly infinite) via a certain nonlinear mapping function of the input data. The wavelet-RX algorithm in this high-dimensional feature space can easily be implemented in terms of kernels that implicitly compute dot products in the feature space (kernelizing the wavelet-RX algorithm). In the proposed kernel wavelet-RX algorithm, a two-dimensional wavelet transform is first applied to decompose the input image into uniform subbands. A number of significant subbands (high-energy subbands) are concatenated together to form a subband-image cube. The kernel RX algorithm is then applied to this subband-image cube. Experimental results are presented for the proposed kernel wavelet-RX, wavelet-RX, and the classical constant false alarm rate (CFAR) algorithm for detecting anomalies (targets) in a large database of LW imagery. The receiver operating characteristic plots show that the proposed kernel wavelet-RX algorithm outperforms the wavelet-RX as well as the classical CFAR detector.

Blind separation of human- and horse-footstep signatures using independent component analysis

Seismic footstep detection based systems for homeland security applications are important to perimeter protection and other security systems. This paper reports seismic footstep signal separation for a walking horse and a walking human. The well-known Independent Component Analysis (ICA) approach is employed to accomplish this task. ICA techniques have become widely used in audio analysis and source separation. The concept of lCA may actually be seen as an extension of the principal component analysis (PCA), which can only impose independence up to the second order and, consequently, defines directions that are orthogonal. They can also be used in conjunction with a classification method to achieve a high percentage of correct classification and reduce false alarms. In this paper, an ICA based algorithm is developed and implemented on seismic data of human and horse footsteps. The performance of this method is very promising and is demonstrated by the experimental results.

Anomaly detection in wavelet domain for long-wave FLIR imagery

This paper describes a new wavelet-based anomaly detection technique for Forward Looking Infrared (FLIR) sensor consisting a Long-wave (LW) and a Mid-wave (MW) sensor. The proposed approach called wavelet-RX algorithm consists of a combination of a two-dimensional (2-D) wavelet transform and the well-known multivariate anomaly detector called the RX algorithm. In our wavelet-RX algorithm, a 2-D wavelet transform is first applied to decompose the input image into uniform subbands. A number of significant subbands (high energy subbands) are concatenated together to form a subband-image cube. The RX algorithm is then applied to each subbandimage cube obtained from wavelet decomposition of LW and MW sensor data separately. Experimental results are presented for the proposed wavelet-RX and the classical CFAR algorithm for detecting anomalies (targets) in a single broadband FLIR (LW or MW) sensors. The results show that the proposed wavelet-RX algorithm outperforms the classical CFAR detector for both LW and for MW FLIR sensors data.

Iterative motion compensation algorithm for heart rate detection with ultrasound data.

Noncontact active ultrasound sensors can be used to detect human physiological signals such as respiration and heart rate using Doppler processing. Target motion such as a person swaying can reduce the performance of these algorithms. To mitigate the effect of target motion or respiration on estimating the heart rate, we developed an iterative motion compensation algorithm. We low pass filter the demodulated data, estimate the motion of the target, and then compensate the data for the slow moving motion. This procedure can be repeated at different cutoff frequencies for different scenarios. Now, standard Doppler processing techniques can be used to analyze the motion compensated data. The algorithm was tested on people standing in a laboratory illuminated with a 40‐KHz continuous‐wave ultrasound sensor. Results for estimating the respiration rate and heart rate will be presented.

Anomaly Detection for Longwave FLIR Imagery Using Kernel Wavelet-RX

This paper describes a new kernel wavelet-based anomaly detection technique for long-wave (LW) Forward Looking Infrared (FLIR) imagery. The proposed approach called kernel wavelet-RX algorithm is essentially an extension of the wavelet-RX algorithm (combination of wavelet transform and RX anomaly detector) to a high dimensional feature space (possibly infinite) via a certain nonlinear mapping function of the input data. The wavelet-RX algorithm in this high dimensional feature space can easily be implemented in terms of kernels that implicitly compute dot products in the feature space (kernelizing the wavelet-RX algorithm). In our kernel wavelet-RX algorithm, a 2-D wavelet transform is first applied to decompose the input image into uniform subbands. A number of significant subbands (high energy subbands) are concatenated together to form a subband-image cube. The kernel RX algorithm is then applied to these subband-image cubes obtained from wavelet decomposition of the LW database images. Experimental results are presented for the proposed kernel wavelet-RX, wavelet-RX and the classical CFAR algorithm for detecting anomalies (targets) in a large database of LW imagery. The ROC plots show that the proposed kernel wavelet-RX algorithm outperforms the wavelet-RX as well as the classical CFAR detector.