Multimodal Biometric Human Recognition for Perceptual Human-Computer Interaction

In this paper, a novel video-based multimodal biometric verification scheme using the subspace-based low-level feature fusion of face and speech is developed for specific speaker recognition for perceptual human--computer interaction (HCI). In the proposed scheme, human face is tracked and face pose is estimated to weight the detected facelike regions in successive frames, where ill-posed faces and false-positive detections are assigned with lower credit to enhance the accuracy. In the audio modality, mel-frequency cepstral coefficients are extracted for voice-based biometric verification. In the fusion step, features from both modalities are projected into nonlinear Laplacian Eigenmap subspace for multimodal speaker recognition and combined at low level. The proposed approach is tested on the video database of ten human subjects, and the results show that the proposed scheme can attain better accuracy in comparison with the conventional multimodal fusion using latent semantic analysis as well as the single-modality verifications. The experiment on MATLAB shows the potential of the proposed scheme to attain the real-time performance for perceptual HCI applications.

Bimodal biometric verification based on face and lips

In this paper, we present a novel approach to person verification by fusing face and lip features. Specifically, the face is modeled by the discriminative common vector and the discrete wavelet transform. Our lip features are simple geometric features based on a lip contour, which can be interpreted as multiple spatial widths and heights from a center of mass. In order to combine these features, we consider two simple fusion strategies: data fusion before training and score fusion after training, working with two different face databases. Fusing them together boosts the performance to achieve an equal error rate as low as 0.4% and 0.28%, respectively, confirming that our approach of fusing lips and face is effective and promising.

Fusion rules for merging uncertain information

In previous papers, we have presented a logic-based framework based on fusion rules for merging structured news reports. Structured news reports are XML documents, where the textentries are restricted to individual words or simple phrases, such as names and domain-specific terminology, and numbers and units. We assume structured news reports do not require natural language processing. Fusion rules are a form of scripting language that define how structured news reports should be merged. The antecedent of a fusion rule is a call to investigate the information in the structured news reports and the background knowledge, and the consequent of a fusion rule is a formula specifying an action to be undertaken to form a merged report. It is expected that a set of fusion rules is defined for any given application. In this paper we extend the approach to handling probability values, degrees of beliefs, or necessity measures associated with textentries in the news reports. We present the formal definition for each of these types of uncertainty and explain how they can be handled using fusion rules. We also discuss the methods of detecting inconsistencies among sources.

Fast heating of ultrahigh-density plasma as a step towards laser fusion ignition

ate studies(2) and fusion energy research(3,4). Laser-driven implosions of spherical polymer shells have, for example, achieved an increase in density of 1,000 times relative to the solid state(5). These densities are large enough to enable controlled fusion, but to achieve energy gain a small volume of compressed fuel (known as the 'spark') must be heated to temperatures of about 10(8) K (corresponding to thermal energies in excess of 10 keV). In the conventional approach to controlled fusion, the spark is both produced and heated by accurately timed shock waves(4), but this process requires both precise implosion symmetry and a very large drive energy. In principle, these requirements can be significantly relaxed by performing the compression and fast heating separately(6-10); however, this 'fast ignitor' approach(7) also suffers drawbacks, such as propagation losses and deflection of the ultra-intense laser pulse by the plasma surrounding the compressed fuel. Here we employ a new compression geometry that eliminates these problems; we combine production of compressed matter in a laser-driven implosion with picosecond-fast heating by a laser pulse timed to coincide with the peak compression. Our approach therefore permits efficient compression and heating to be carried out simultaneously, providing a route to efficient fusion energy production.

Fast Heating scalable to laser fusion ignition

Rapid heating of a compressed fusion fuel by a short-duration laser pulse is a promising route to generating energy by nuclear fusion1, and has been demonstrated on an experimental scale using a novel fast-ignitor geometry2. Here we describe a refinement of this system in which a much more powerful, pulsed petawatt (1015 watts) laser creates a fastheated core plasma that is scalable to fullscale ignition, significantly increasing the number of fusion events while still maintaining high heating efficiency at these substantially higher laser energies. Our findings bring us a step closer to realizing the production of relatively inexpensive, full-scale fast-ignition laser facilities.

FPGA-based minutia matching for biometric fingerprint image database retrieval

In this paper, a parallel-matching processor architecture with early jump-out (EJO) control is proposed to carry out high-speed biometric fingerprint database retrieval. The processor performs the fingerprint retrieval by using minutia point matching. An EJO method is applied to the proposed architecture to speed up the large database retrieval. The processor is implemented on a Xilinx Virtex-E, and occupies 6,825 slices and runs at up to 65 MHz. The software/hardware co-simulation benchmark with a database of 10,000 fingerprints verifies that the matching speed can achieve the rate of up to 1.22 million fingerprints per second. EJO results in about a 22% gain in computing efficiency.