Hoda Mohammadzade

Iterative Closest Normal Point for 3D Face Recognition

The common approach for 3D face recognition is to register a probe face to each of the gallery faces and then calculate the sum of the distances between their points. This approach is computationally expensive and sensitive to facial expression variation. In this paper, we introduce the iterative closest normal point method for finding the corresponding points between a generic reference face and every input face. The proposed correspondence finding method samples a set of points for each face, denoted as the closest normal points. These points are effectively aligned across all faces, enabling effective application of discriminant analysis methods for 3D face recognition. As a result, the expression variation problem is addressed by minimizing the within-class variability of the face samples while maximizing the between-class variability. As an important conclusion, we show that the surface normal vectors of the face at the sampled points contain more discriminatory information than the coordinates of the points. We have performed comprehensive experiments on the Face Recognition Grand Challenge database, which is presently the largest available 3D face database. We have achieved verification rates of 99.6 and 99.2 percent at a false acceptance rate of 0.1 percent for the all versus all and ROC III experiments, respectively, which, to the best of our knowledge, have seven and four times less error rates, respectively, compared to the best existing methods on this database.

Projection into Expression Subspaces for Face Recognition from Single Sample per Person

Discriminant analysis methods are powerful tools for face recognition. However, these methods cannot be used for the single sample per person scenario because the within-subject variability cannot be estimated in this case. In the generic learning solution, this variability is estimated using images of a generic training set, for which more than one sample per person is available. However, because of rather poor estimation of the within-subject variability using a generic set, the performance of discriminant analysis methods is yet to be satisfactory. This problem particularly exists when images are under drastic facial expression variation. In this paper, we show that images with the same expression are located on a common subspace, which here we call it the expression subspace. We show that by projecting an image with an arbitrary expression into the expression subspaces, we can synthesize new expression images. By means of the synthesized images for subjects with one image sample, we can obtain more accurate estimation of the within-subject variability and achieve significant improvement in recognition. We performed comprehensive experiments on two large face databases: the Face Recognition Grand Challenge and the Cohn-Kanade AU-Coded Facial Expression database to support the proposed methodology.

Cuff-less high-accuracy calibration-free blood pressure estimation using pulse transit time

Recently a few methods have been proposed in the literature for non-invasive cuff-less estimation of systolic and diastolic blood pressures. One of the most prominent methods is to use the Pulse Transit Time (PTT). Although it is proven that PTT has a strong correlation with the systolic and diastolic blood pressures, this relation is highly dependent to each individuals physiological properties. Therefore, it requires per person calibration for accurate and reliable blood pressure estimation from PTT, which is a big drawback. To alleviate this issue, in this paper, a novel method is proposed for accurate and reliable estimation of blood pressure that is calibration-free. This goal is accomplished by extraction of several physiological parameters from Photoplethysmography (PPG) signal as well as utilizing signal processing and machine learning algorithms. The results show that the accuracy of the proposed method achieves grade B for the estimation of the diastolic blood pressure and grade C for the estimation of the mean arterial pressure under the standard British Hypertension Society (BHS) protocol.

Cuffless Blood Pressure Estimation Algorithms for Continuous Health-Care Monitoring

Goal: Continuous blood pressure (BP) monitoring can provide invaluable information about individuals’ health conditions. However, BP is conventionally measured using inconvenient cuff-based instruments, which prevents continuous BP monitoring. This paper presents an efficient algorithm, based on the pulse arrival time (PAT), for the continuous and cuffless estimation of the systolic BP, diastolic blood pressure (DBP), and mean arterial pressure (MAP) values. Methods: The proposed framework estimates the BP values through processing vital signals and extracting two types of features, which are based on either physiological parameters or whole-based representation of vital signals. Finally, the regression algorithms are employed for the BP estimation. Although the proposed algorithm works reliably without any need for calibration, an optional calibration procedure is also suggested, which can improve the systems accuracy even further. Results: The proposed method is evaluated on about a thousand subjects using the Association for the Advancement of Medical Instrumentation (AAMI) and the British Hypertension Society (BHS) standards. The method complies with the AAMI standard in the estimation of DBP and MAP values. Regarding the BHS protocol, the results achieve grade A for the estimation of DBP and grade B for the estimation of MAP. Conclusion: We conclude that by using the PAT in combination with informative features from the vital signals, the BP can be accurately and reliably estimated in a noninvasive fashion. Significance: The results indicate that the proposed algorithm for the cuffless estimation of the BP can potentially enable mobile health-care gadgets to monitor the BP continuously.

An expression transformation for improving the recognition of expression-variant faces from one sample image per person

It is known that when only one sample image per gallery person is available, as a result of the small sample size problem, the recognition performance of discriminant feature extraction methods substantially degrades. This is particularly the case when the images are under drastic facial expression variation. To address this problem, this paper introduces an appearance-based expression transformation method to synthesize new expression images from the probe image. By feeding the synthesized images to discriminant feature extractor, a more robust recognition of gallery images with single sample image is achieved. The effectiveness of the proposed transformation method is demonstrated using the Cohn-Kanade facial expression database.

ECG for blind identity verification in distributed systems

This paper discusses ECG biometric recognition in a distributed system, such as smart cards. In a setting where every card is equipped with an ECG sensor to record heart beats from the fingers, and to subsequently perform identity verification, the interest is in protecting the card holder from a set of unknown/unseen biometric traits. Prior works have examined ECG biometrics in settings where a particular subject was to be identified among a set of enrollees. However, this treatment limits the applicability of this biometric. The Autocorrelation - Linear Discriminant Analysis (AC/LDA) is revisited, to propose a strategic extension of the methodology, in order to account for recognition among unknown individuals (blind verification). The discriminant is trained individually for every smart card, on the samples of the subject to be enrolled, as well as a generic dataset of ECG recordings. This enables the recognizer to protect the template against attacks by biometric samples that have not been used to train the discriminant. In addition, we present a methodology for the selection of the matching threshold, which targets to control false acceptance while being experimentally optimized for a particular smart card.

A Simultaneous Div-Curl 2D Clifford Fourier Transform Filter for Enhancing Vortices, Sinks and Sources in Sampled 2D Vector Field Images

A novel two-dimensional (2D) filtering operation, involving both curl and divergence, is applied to the 2D Clifford Fourier transform (CFT) in order to simultaneously enhance important features of a 2D vector field, such as vortices and pairs of sources and sinks. Applications are envisaged for enhancing vector field images in such sampled vector field images as fluid flow, gas flow and biomedical image processing.

BEMD for expression transformation in face recognition

This work presents a novel methodology for the transformation of facial expressions, to assist face biometrics. It is known that identification using only one image per subject poses a great challenge to recognizers. This is because drastic facial expressions introduce variability, on which the recognizer is not trained. The proposed framework uses only one image per subject to predict intra-class variability, by synthesizing new expressions, which are subsequently used to train the discriminant. The expression of the gallery is transformed using the bivariate empirical mode decomposition (BEMD), which allows for simultaneous analysis of the probe image and a targeted expression mask. We advocate that 2D BEMD is a powerful tool for multi-resolution face analysis. The performance of the proposed framework, tested over a database of 96 individuals, is 90% for an FAR of 1%.

A 2D bivariate EMD algorithm for image fusion

Although the benefits of the empirical mode decomposition, in analyzing stochastic signals have been reported, and the algorithm has been established for fusion applications, there is currently no solution to the problem of the simultaneous decomposition of 2D data. This paper proposes an extension of the bivariate EMD (BEMD) [1] algorithm for 2D sources, which retains spatial information while addressing the uniqueness problem of the intrinsic mode functions. The performance of the algorithm is tested on partially blurred and defocused images. The fused images are compared against 1D-BEMD solutions, to demonstrate increased visual quality of the result.

Critical object recognition in millimeter-wave images with robustness to rotation and scale

Locating critical objects is crucial in various security applications and industries. For example, in security applications, such as in airports, these objects might be hidden or covered under shields or secret sheaths. Millimeter-wave images can be utilized to discover and recognize the critical objects out of the hidden cases without any health risk due to their non-ionizing features. However, millimeter-wave images usually have waves in and around the detected objects, making object recognition difficult. Thus, regular image processing and classification methods cannot be used for these images and additional pre-processings and classification methods should be introduced. This paper proposes a novel pre-processing method for canceling rotation and scale using principal component analysis. In addition, a two-layer classification method is introduced and utilized for recognition. Moreover, a large dataset of millimeter-wave images is collected and created for experiments. Experimental results show that a typical classification method such as support vector machines can recognize 45.5% of a type of critical objects at 34.2% false alarm rate (FAR), which is a drastically poor recognition. The same method within the proposed recognition framework achieves 92.9% recognition rate at 0.43% FAR, which indicates a highly significant improvement. The significant contribution of this work is to introduce a new method for analyzing millimeter-wave images based on machine vision and learning approaches, which is not yet widely noted in the field of millimeter-wave image analysis.