Distance-driven fusion of gait and face for human identification in video

Gait and face are two important biometrics for human identification. Complementary properties of these two biometrics suggest fusion of them. The relationship between gait and face in the fusion is affected by the subject-to-camera distance. On the one hand, gait is a suitable biometric trait for human recognition at a distance. On the other hand, face recognition is more reliable when the subject is close to the camera. This paper proposes an adaptive fusion method called distance-driven fusion to combine gait and face for human identification in video. Rather than predefined fixed fusion rules, distance-driven fusion dynamically adjusts its rule according to the subject-to-camera distance in real time. Experimental results show that distance-driven fusion performs better than not only single biometric, but also the conventional
static fusion rules including MEAN, PRODUCT, MIN, and MAX.

Adaptive fusion of gait and face for human identification in video

Most work on multi-biometric fusion is based on static fusion rules which cannot respond to the changes of the environment and the individual users. This paper proposes adaptive multi-biometric fusion, which dynamically adjusts the fusion rules to suit the real-time external conditions. As a typical example, the adaptive fusion of gait and face in video is studied. Two factors that may affect the relationship between gait and face in the fusion are considered, i.e., the view angle and the subject-to-camera distance. Together they determine the way gait and face are fused at an arbitrary time. Experimental results show that the adaptive fusion performs significantly better than not only single biometric traits, but also those widely adopted static fusion rules including SUM, PRODUCT, MIN, and MAX.

K-Complex Detection Using a Hybrid-Synergic Machine Learning Method

Sleep stage identification is the first step in modern sleep disorder diagnostics process. K-complex is an indicator for the sleep stage 2. However, due to the ambiguity of the translation of the medical standards into a computer-based procedure, reliability of automated K-complex detection from the EEG wave is still far from expectation. More specifically, there are some significant barriers to the research of automatic K-complex detection. First, there is no adequate description of K-complex that makes it difficult to develop automatic detection algorithm. Second, human experts only provided the label for whether a whole EEG segment contains K-complex or not, rather than individual labels for each subsegment. These barriers render most pattern recognition algorithms inapplicable in detecting K-complex. In this paper, we attempt to address these two challenges, by designing a new feature extraction method that can transform visual features of the EEG wave with any length into mathematical representation and proposing a hybrid-synergic machine learning method to build a K-complex classifier. The tenfold cross-validation results indicate that both the accuracy and the precision of this proposed model are at least as good as a human expert in K-complex detection.

Multi-classifier classification of spam email on an ubiquitous multi-core architecture

This paper presents an innovative fusion based multi-classifier email classification on a ubiquitous multi-core architecture. Many approaches use text-based single classifiers or multiple weakly trained classifiers to identify spam messages from a large email corpus. We build upon our previous work on multi-core by apply our ubiquitous multi-core framework to run our fusion based multi-classifier architecture. By running each classifier process in parallel within their dedicated core, we greatly improve the performance of our proposed multi-classifier based filtering system. Our proposed architecture also provides a safeguard of user mailbox from different malicious attacks. Our experimental results show that we achieved an average of 30% speedup at the average cost of 1.4 ms. We also reduced the instance of false positive, which is one of the key challenges in spam filtering system, and increases email classification accuracy substantially compared with single classification techniques.

Robust multi-sensor data fusion

This research obtains the optimal estimation and data fusion for linear and nonlinear systems suffering from uncertain observations (missing measurements). The noise from the different data sources are considered to correlated. The derivation of the robust Kalman filter for systems subject to aditional uncertainties in the modelling parameters is presented.

Multi-sensory face biometric fusion (for personal identification)

The objective of this work is to recognize faces using sets of images in visual and thermal spectra. This is challenging because the former is greatly affected by illumination changes, while the latter frequently contains occlusions due to eye-wear and is inherently less discriminative. Our method is based on a fusion of the two modalities. Specifically: we examine (i) the effects of preprocessing of data in each domain, (ii) the fusion of holistic and local facial appearance, and (iii) propose an algorithm for combining the similarity scores in visual and thermal spectra in the presence of prescription glasses and significant pose variations, using a small number of training images (5-7). Our system achieved a high correct identification rate of 97% on a freely available test set of 29 individuals and extreme illumination changes.

Structural damage identification based on self-fitting ARMAX model and multi-sensor data fusion

Statistical time series methods have proven to be a promising technique in structural health monitoring, since it provides a direct form of data analysis and eliminates the requirement for domain transformation. Latest research in structural health monitoring presents a number of statistical models that have been successfully used to construct quantified models of vibration response signals. Although a majority of these studies present viable results, the aspects of practical implementation, statistical model construction and decision-making procedures are often vaguely defined or omitted from presented work. In this article, a comprehensive methodology is developed, which essentially utilizes an auto-regressive moving average with exogenous input model to create quantified model estimates of experimentally acquired response signals. An iterative self-fitting algorithm is proposed to construct and fit the auto-regressive moving average with exogenous input model, which is capable of integrally finding an optimum set of auto-regressive moving average with exogenous input model parameters. After creating a dataset of quantified response signals, an unlabelled response signal can be identified according to a 'closest-fit' available in the dataset. A unique averaging method is proposed and implemented for multi-sensor data fusion to decrease the margin of error with sensors, thus increasing the reliability of global damage identification. To demonstrate the effectiveness of the developed methodology, a steel frame structure subjected to various bolt-connection damage scenarios is tested. Damage identification results from the experimental study suggest that the proposed methodology can be employed as an efficient and functional damage identification tool. © The Author(s) 2014.

Multi-Kinect skeleton fusion for physical rehabilitation monitoring

 Kinect has been increasingly applied in rehabilitation as a motion capture device. However, the inherent limitations significantly hinder its further development in this important area. Although a number of Kinect fusion approaches have been proposed, only a few of them was actually considered for rehabilitation. In this paper, we propose to fuse information from multiple Kinects to achieve this. Given the specific scenario of users suffering from limited range of movements, we propose to calibrate depth cameras in multiple Kinects with 3D positions of joints on a human body rather than in a checkerboard pattern, so that patients are able to calibrate Kinects without extra support. Kalman filter is applied for skeleton-wise Kinect fusion since skeleton data (3D positions of joints) and its derivatives are preferred by physiotherapists to evaluate the exercise performance of patients. Various preliminary experiments were conducted to illustrate the accuracy of proposed calibration and fusion approach by comparing with a commercial Vicon system®, confirming the practical use of the system in rehabilitation exercise monitoring.

Toxicity prediction in cancer using multiple instance learning in a multi-task framework

Treatments of cancer cause severe side effects called toxicities. Reduction of such effects is crucial in cancer care. To impact care, we need to predict toxicities at fortnightly intervals. This toxicity data differs from traditional time series data as toxicities can be caused by one treatment on a given day alone, and thus it is necessary to consider the effect of the singular data vector causing toxicity. We model the data before prediction points using the multiple instance learning, where each bag is composed of multiple instances associated with daily treatments and patient-specific attributes, such as chemotherapy, radiotherapy, age and cancer types. We then formulate a Bayesian multi-task framework to enhance toxicity prediction at each prediction point. The use of the prior allows factors to be shared across task predictors. Our proposed method simultaneously captures the heterogeneity of daily treatments and performs toxicity prediction at different prediction points. Our method was evaluated on a real-word dataset of more than 2000 cancer patients and had achieved a better prediction accuracy in terms of AUC than the state-of-art baselines.

Sensor fusion based localization of a mobile user in a wireless network

This paper applies sensor fusion to the localization problem of a mobile user. We propose that the use of direction of arrival (DOA) estimations along with received signal strength measurements can increase the accuracy and robustness of location estimations. The DOA estimations are incapable of providing multi-dimensional positioning alone, while signal strength methods are prone to high uncertainties. A Robust Extended Kalman Filter (REKF) is used to derive the state estimate of the mobile user's position, and successfully track the mobile users with less system complexity, as it requires measurements from only one base station. Therefore, localization of mobile users can be performed at the single base station. Furthermore, the technique is robust against system uncertainties caused by the inherent deterministic nature of the mobility model. Through simulation, we show the accuracy of our prediction algorithm and the simplicity of its implementation.

Dynamic biometrics fusion at feature level for video-based human recognition

This paper proposes a novel human recognition method in video, which combines human face and gait traits
using a dynamic multi-modal biometrics fusion scheme. The Fisherface approach is adopted to extract face
features, while for gait features, Locality Preserving Projection (LPP) is used to achieve low-dimensional
manifold embedding of the temporal silhouette data derived from image sequences. Face and gait features are
fused dynamically at feature level based on a distance-driven fusion method. Encouraging experimental results
are achieved on the video sequences containing 20 people, which show that dynamically fused features produce
a more discriminating power than any individual biometric as well as integrated features built on common static
fusion schemes.

Generation of mutant leukaemia inhibitory factor (LIF)-IgG heavy chain fusion proteins as bivalent antagonists of LIF

Two leukaemia inhibitory factor (LIF) mutants, designated MH35-BD and LIF05, have been shown to have a capacity to inhibit the biological activities of not only human LIF (hLIF) but also other interleukin-6 (IL-6) subfamily cytokines such as human oncostatin M (hOSM). These cytokines share the same receptor complex in which the glycoprotein 130 (gp130) subunit is a common constituent. However, at low concentrations and in their monomeric forms, such molecules have a relatively short plasma half-life due to rapid clearance from the kidneys. Here, to prolong their serum half-lives, we have used a multi-step polymerase chain reaction (PCR) to fuse each of the LIF05 and MH35-BD cDNA fragments to a sequence encoding the Fc portion, and the hinge region, of the human immunoglobulin G (hIgG) heavy chain. The linking was achieved through an oligomer encoding a thrombin-sensitive peptide linker thus generating MH35-BD:Fc and LIF05:Fc, respectively. Both Fc fusion constructs were expressed in insect cell Sf21 and the proteins were purified by two successive affinity chromatography steps using nickel–nitrilotriacetic acid (Ni–NTA) agarose and protein A beads. The Ba/F3 cell-based proliferation assay was used to confirm that the proteins were biologically active. In addition, preliminary pharmacokinetics indicates that the Fc fusion constructs have a longer serum half-life compared to their non-fusion counterparts.

Knowledge discovery in multi-databases by analyzing local instances

Multi-databases mining is an urgent task. This thesis solves 4 key problems in multi-databases mining: Application-independent database classification - Local instance analysis model - Useful pattern discovery - Pattern synthesis.

Spam filtering for network traffic security on a multi-core environment

This paper presents an innovative fusion-based multi-classifier e-mail classification on a ubiquitous multicore architecture. Many previous approaches used text-based single classifiers to identify spam messages from a large e-mail corpus with some amount of false positive tradeoffs. Researchers are trying to prevent false positive in their filtering methods, but so far none of the current research has claimed zero false positive results. In e-mail classification false positive can potentially cause serious problems for the user. In this paper, we use fusion-based multi-classifier classification technique in a multi-core framework. By running each classifier process in parallel within their dedicated core, we greatly improve the performance of our multi-classifier-based filtering system in terms of running time, false positive rate, and filtering accuracy. Our proposed architecture also provides a safeguard of user mailbox from different malicious attacks. Our experimental results show that we achieved an average of 30% speedup at an average cost of 1.4 ms. We also reduced the instances of false positives, which are one of the key challenges in a spam filtering system, and increases e-mail classification accuracy substantially compared with single classification techniques.