Classifier selection using sequential error ratio criterion for multi-instance and multi-sample fusion

Classifier selection is a problem encountered by multi-biometric systems that aim to improve performance through fusion of decisions. A particular decision fusion architecture that combines multiple instances (n classifiers) and multiple samples (m attempts at each classifier) has been proposed in previous work to achieve controlled trade-off between false alarms and false rejects. Although analysis on text-dependent speaker verification has demonstrated better performance for fusion of decisions with favourable dependence compared to statistically independent decisions, the performance is not always optimal. Given a pool of instances, best performance with this architecture is obtained for certain combination of instances. Heuristic rules and diversity measures have been commonly used for classifier selection but it is shown that optimal performance is achieved for the `best combination performance' rule. As the search complexity for this rule increases exponentially with the addition of classifiers, a measure - the sequential error ratio (SER) - is proposed in this work that is specifically adapted to the characteristics of sequential fusion architecture. The proposed measure can be used to select a classifier that is most likely to produce a correct decision at each stage. Error rates for fusion of text-dependent HMM based speaker models using SER are compared with other classifier selection methodologies. SER is shown to achieve near optimal performance for sequential fusion of multiple instances with or without the use of multiple samples. The methodology applies to multiple speech utterances for telephone or internet based access control and to other systems such as multiple finger print and multiple handwriting sample based identity verification systems.

Sequential decision fusion of multibiometrics applied to text-dependent speaker verification for controlled errors

Reliability of the performance of biometric identity verification systems remains a significant challenge. Individual biometric samples of the same person (identity class) are not identical at each presentation and performance degradation arises from intra-class variability and inter-class similarity. These limitations lead to false accepts and false rejects that are dependent. It is therefore difficult to reduce the rate of one type of error without increasing the other. The focus of this dissertation is to investigate a method based on classifier fusion techniques to better control the trade-off between the verification errors using text-dependent speaker verification as the test platform. A sequential classifier fusion architecture that integrates multi-instance and multisample fusion schemes is proposed. This fusion method enables a controlled trade-off between false alarms and false rejects. For statistically independent classifier decisions, analytical expressions for each type of verification error are derived using base classifier performances. As this assumption may not be always valid, these expressions are modified to incorporate the correlation between statistically dependent decisions from clients and impostors. The architecture is empirically evaluated by applying the proposed architecture for text dependent speaker verification using the Hidden Markov Model based digit dependent speaker models in each stage with multiple attempts for each digit utterance. The trade-off between the verification errors is controlled using the parameters, number of decision stages (instances) and the number of attempts at each decision stage (samples), fine-tuned on evaluation/tune set. The statistical validation of the derived expressions for error estimates is evaluated on test data. The performance of the sequential method is further demonstrated to depend on the order of the combination of digits (instances) and the nature of repetitive attempts (samples). The false rejection and false acceptance rates for proposed fusion are estimated using the base classifier performances, the variance in correlation between classifier decisions and the sequence of classifiers with favourable dependence selected using the 'Sequential Error Ratio' criteria. The error rates are better estimated by incorporating user-dependent (such as speaker-dependent thresholds and speaker-specific digit combinations) and class-dependent (such as clientimpostor dependent favourable combinations and class-error based threshold estimation) information. The proposed architecture is desirable in most of the speaker verification applications such as remote authentication, telephone and internet shopping applications. The tuning of parameters - the number of instances and samples - serve both the security and user convenience requirements of speaker-specific verification. The architecture investigated here is applicable to verification using other biometric modalities such as handwriting, fingerprints and key strokes.

Channel estimation relying on the minimum bit-error-ratio criterion for BPSK and QPSK signals

The authors consider the channel estimation problem in the context of a linear equaliser designed for a frequency selective channel, which relies on the minimum bit-error-ratio (MBER) optimisation framework. Previous literature has shown that the MBER-based signal detection may outperform its minimum-mean-square-error (MMSE) counterpart in the bit-error-ratio performance sense. In this study, they develop a framework for channel estimation by first discretising the parameter space and then posing it as a detection problem. Explicitly, the MBER cost function (CF) is derived and its performance studied, when transmitting binary phase shift keying (BPSK) and quadrature phase shift keying (QPSK) signals. It is demonstrated that the MBER based CF aided scheme is capable of outperforming existing MMSE, least square-based solutions.

Communication-efficient decentralized sequential detection

The problem of decentralized sequential detection is studied in this thesis, where local sensors are memoryless, receive independent observations, and no feedback from the fusion center. In addition to traditional criteria of detection delay and error probability, we introduce a new constraint: the number of communications between local sensors and the fusion center. This metric is able to reflect both the cost of establishing communication links as well as overall energy consumption over time. A new formulation for communication-efficient decentralized sequential detection is proposed where the overall detection delay is minimized with constraints on both error probabilities and the communication cost. Two types of problems are investigated based on the communication-efficient formulation: decentralized hypothesis testing and decentralized change detection. In the former case, an asymptotically person-by-person optimum detection framework is developed, where the fusion center performs a sequential probability ratio test based on dependent observations. The proposed algorithm utilizes not only reported statistics from local sensors, but also the reporting times. The asymptotically relative efficiency of proposed algorithm with respect to the centralized strategy is expressed in closed form. When the probabilities of false alarm and missed detection are close to one another, a reduced-complexity algorithm is proposed based on a Poisson arrival approximation. In addition, decentralized change detection with a communication cost constraint is also investigated. A person-by-person optimum change detection algorithm is proposed, where transmissions of sensing reports are modeled as a Poisson process. The optimum threshold value is obtained through dynamic programming. An alternative method with a simpler fusion rule is also proposed, where the threshold values in the algorithm are determined by a combination of sequential detection analysis and constrained optimization. In both decentralized hypothesis testing and change detection problems, tradeoffs in parameter choices are investigated through Monte Carlo simulations.

Cooperative distributed sequential spectrum sensing

We consider cooperative spectrum sensing for cognitive radios. We develop an energy efficient detector with low detection delay using sequential hypothesis testing. Sequential Probability Ratio Test (SPRT) is used at both the local nodes and the fusion center. We also analyse the performance of this algorithm and compare with the simulations. Modelling uncertainties in the distribution parameters are considered. Slow fading with and without perfect channel state information at the cognitive radios is taken into account.

A novel algorithm for cooperative distributed sequential spectrum sensing in Cognitive Radio

This paper considers cooperative spectrum sensing in Cognitive Radios. In our previous work we have developed DualSPRT, a distributed algorithm for cooperative spectrum sensing using Sequential Probability Ratio Test (SPRT) at the Cognitive Radios as well as at the fusion center. This algorithm works well, but is not optimal. In this paper we propose an improved algorithm- SPRT-CSPRT, which is motivated from Cumulative Sum Procedures (CUSUM). We analyse it theoretically. We also modify this algorithm to handle uncertainties in SNR's and fading.

Cooperative distributed sequential spectrum sensing

We consider cooperative spectrum sensing for cognitive radios. We develop an energy efficient detector with low detection delay using sequential hypothesis testing. Sequential Probability Ratio Test (SPRT) is used at both the local nodes and the fusion center. We also analyse the performance of this algorithm and compare with the simulations. Modelling uncertainties in the distribution parameters are considered. Slow fading with and without perfect channel state information at the cognitive radios is taken into account.

Spectrum sensing using distributed sequential detection via noisy reporting MAC

This paper considers cooperative spectrum sensing algorithms for Cognitive Radios which focus on reducing the number of samples to make a reliable detection. We propose algorithms based on decentralized sequential hypothesis testing in which the Cognitive Radios sequentially collect the observations, make local decisions and send them to the fusion center for further processing to make a final decision on spectrum usage. The reporting channel between the Cognitive Radios and the fusion center is assumed more realistically as a Multiple Access Channel (MAC) with receiver noise. Furthermore the communication for reporting is limited, thereby reducing the communication cost. We start with an algorithm where the fusion center uses an SPRT-like (Sequential Probability Ratio Test) procedure and theoretically analyze its performance. Asymptotically, its performance is close to the optimal centralized test without fusion center noise. We further modify this algorithm to improve its performance at practical operating points. Later we generalize these algorithms to handle uncertainties in SNR and fading. (C) 2014 Elsevier B.V. All rights reserved.

Minimum-Error-Probability CFO Estimation for Multiuser MIMO-OFDM Systems

We consider carrier frequency offset (CFO) estimation in the context of multiple-input multiple-output (MIMO) orthogonal frequency-division multiplexing (OFDM) systems over noisy frequency-selective wireless channels with both single- and multiuser scenarios. We conceived a new approach for parameter estimation by discretizing the continuous-valued CFO parameter into a discrete set of bins and then invoked detection theory, analogous to the minimum-bit-error-ratio optimization framework for detecting the finite-alphabet received signal. Using this radical approach, we propose a novel CFO estimation method and study its performance using both analytical results and Monte Carlo simulations. We obtain expressions for the variance of the CFO estimation error and the resultant BER degradation with the single- user scenario. Our simulations demonstrate that the overall BER performance of a MIMO-OFDM system using the proposed method is substantially improved for all the modulation schemes considered, albeit this is achieved at increased complexity.

Attending to risk in sequential sampling plans

Researchers typically tackle questions by constructing powerful, highlyreplicated sampling protocols or experimental designs. Such approaches often demand large samples sizes and are usually only conducted on a once-off basis. In contrast, many industries need to continually monitor phenomena such as equipment reliability, water quality, or the abundance of a pest. In such instances, costs and time inherent in sampling preclude the use of highlyintensive methods. Ideally, one wants to collect the absolute minimum number of samples needed to make an appropriate decision. Sequential sampling, wherein the sample size is a function of the results of the sampling process itself, offers a practicable solution. But smaller sample sizes equate to less knowledge about the population, and thus an increased risk of making an incorrect management decision. There are various statistical techniques to account for and measure risk in sequential sampling plans. We illustrate these methods and assess them using examples relating to the management of arthropod pests in commercial crops, but they can be applied to any situation where sequential sampling is used.

Planos de amostragem seqüenciais para Orthezia praelonga Douglas (Hemiptera: Sternorrhyncha, Ortheziidae) na cultura de citros

Coordenação de Aperfeiçoamento de Pessoal de Nível Superior (CAPES)

Sequential sampling of Euschistus heros (Heteroptera: Pentatomidae) in soybean

Coordenação de Aperfeiçoamento de Pessoal de Nível Superior (CAPES)

Application of commercially available, low-cost, miniaturised NIR spectrometers to the assessment of the sugar content of intact fruit

Recent decreases in costs, and improvements in performance, of silicon array detectors open a range of potential applications of relevance to plant physiologists, associated with spectral analysis in the visible and short-wave near infra-red (far-red) spectrum. The performance characteristics of three commercially available ‘miniature’ spectrometers based on silicon array detectors operating in the 650–1050-nm spectral region (MMS1 from Zeiss, S2000 from Ocean Optics, and FICS from Oriel, operated with a Larry detector) were compared with respect to the application of non-invasive prediction of sugar content of fruit using near infra-red spectroscopy (NIRS). The FICS–Larry gave the best wavelength resolution; however, the narrow slit and small pixel size of the charge-coupled device detector resulted in a very low sensitivity, and this instrumentation was not considered further. Wavelength resolution was poor with the MMS1 relative to the S2000 (e.g. full width at half maximum of the 912 nm Hg peak, 13 and 2 nm for the MMS1 and S2000, respectively), but the large pixel height of the array used in the MMS1 gave it sensitivity comparable to the S2000. The signal-to-signal standard error ratio of spectra was greater by an order of magnitude with the MMS1, relative to the S2000, at both near saturation and low light levels. Calibrations were developed using reflectance spectra of filter paper soaked in range of concentrations (0–20% w/v) of sucrose, using a modified partial least squares procedure. Calibrations developed with the MMS1 were superior to those developed using the S2000 (e.g. coefficient of correlation of 0.90 and 0.62, and standard error of cross-validation of 1.9 and 5.4%, respectively), indicating the importance of high signal to noise ratio over wavelength resolution to calibration accuracy. The design of a bench top assembly using the MMS1 for the non-invasive assessment of mesocarp sugar content of (intact) melon fruit is reported in terms of light source and angle between detector and light source, and optimisation of math treatment (derivative condition and smoothing function).

On the performance analysis of a class of neuron circuits

A new performance metric, Peak-Error Ratio (PER) has been presented to benchmark the performance of a class of neuron circuits to realize neuron activation function (NAF) and its derivative (DNAF). Neuron circuits, biased in subthreshold region, based on the asymmetric cross-coupled differential pair configuration and conventional configuration of applying small external offset voltage at the input have been compared on the basis of PER. It is shown that the technique of using transistor asymmetry in a cross-coupled differential pair performs on-par with that of applying external offset voltage. The neuron circuits have been experimentally prototyped and characterized as a proof of concept on the 1.5 mu m AMI technology.

On the performance analysis of a class of neuron circuits

A new performance metric, Peak-Error Ratio (PER) has been presented to benchmark the performance of a class of neuron circuits to realize neuron activation function (NAF) and its derivative (DNAF). Neuron circuits, biased in subthreshold region, based on the asymmetric cross-coupled differential pair configuration and conventional configuration of applying small external offset voltage at the input have been compared on the basis of PER. It is shown that the technique of using transistor asymmetry in a cross-coupled differential pair performs on-par with that of applying external offset voltage. The neuron circuits have been experimentally prototyped and characterized as a proof of concept on the 1.5 mu m AMI technology.