Local Decision Thresholds Research Articles

Person Re-Identification With Metric Learning Using Privileged Information

Despite the promising progress made in recent years, person re-identification remains a challenging task due to complex variations in human appearances from different camera views. This paper presents a logistic discriminant metric learning method for this challenging problem. Different with most existing metric learning algorithms, it exploits both original data and auxiliary data during training, which is motivated by the new machine learning paradigm-learning using privileged information. Such privileged information is a kind of auxiliary knowledge, which is only available during training. Our goal is to learn an optimal distance function by constructing a locally adaptive decision rule with the help of privileged information. We jointly learn two distance metrics by minimizing the empirical loss penalizing the difference between the distance in the original space and that in the privileged space. In our setting, the distance in the privileged space functions as a local decision threshold, which guides the decision making in the original space like a teacher. The metric learned from the original space is used to compute the distance between a probe image and a gallery image during testing. In addition, we extend the proposed approach to a multi-view setting which is able to explore the complementation of multiple feature representations. In the multi-view setting, multiple metrics corresponding to different original features are jointly learned, guided by the same privileged information. Besides, an effective iterative optimization scheme is introduced to simultaneously optimize the metrics and the assigned metric weights. Experiment results on several widely-used data sets demonstrate that the proposed approach is superior to global decision threshold-based methods and outperforms most state-of-the-art results.

Secure Distributed Detection under Energy Constraint in IoT-Oriented Sensor Networks.

We study the secure distributed detection problems under energy constraint for IoT-oriented sensor networks. The conventional channel-aware encryption (CAE) is an efficient physical-layer secure distributed detection scheme in light of its energy efficiency, good scalability and robustness over diverse eavesdropping scenarios. However, in the CAE scheme, it remains an open problem of how to optimize the key thresholds for the estimated channel gain, which are used to determine the sensor’s reporting action. Moreover, the CAE scheme does not jointly consider the accuracy of local detection results in determining whether to stay dormant for a sensor. To solve these problems, we first analyze the error probability and derive the optimal thresholds in the CAE scheme under a specified energy constraint. These results build a convenient mathematic framework for our further innovative design. Under this framework, we propose a hybrid secure distributed detection scheme. Our proposal can satisfy the energy constraint by keeping some sensors inactive according to the local detection confidence level, which is characterized by likelihood ratio. In the meanwhile, the security is guaranteed through randomly flipping the local decisions forwarded to the fusion center based on the channel amplitude. We further optimize the key parameters of our hybrid scheme, including two local decision thresholds and one channel comparison threshold. Performance evaluation results demonstrate that our hybrid scheme outperforms the CAE under stringent energy constraints, especially in the high signal-to-noise ratio scenario, while the security is still assured.

Genotyping for Human Papillomavirus Types 16 and 18 in Women With Minor Cervical Lesions: A Systematic Review and Meta-analysis.

High-risk human papillomavirus (hrHPV) testing to triage women with minor cervical lesions generates many referrals. To evaluate the accuracy of genotyping for HPV types 16 and 18 and its utility as a second triage step after hrHPV testing in women with minor cervical lesions. Searches of 4 bibliographic databases, without language restrictions, from 1 January 1999 to 1 February 2016. Studies involving women with atypical squamous cells of undetermined significance (ASC-US) or low-grade squamous intraepithelial lesions (LSIL) who were triaged with tests for hrHPV and HPV 16/18 to find cervical intraepithelial neoplasia (grade ≥2 [CIN2+] or grade ≥3 [CIN3+]). Independent study selection, extraction of data, and quality assessment by 2 reviewers. Twenty-four moderate- to good-quality studies involving 8587 women with ASC-US and 5284 with LSIL were found. The pooled sensitivity of HPV 16/18 genotyping for CIN3+ was about 70% for women with either ASC-US or LSIL. The pooled specificity (with a threshold of grade <2 CIN) was 83% (95% CI, 80% to 86%) for women with ASC-US and 76% (CI, 74% to 79%) for those with LSIL. The average risk for CIN3+ was 17% and 19% in HPV 16/18-positive women with ASC-US and LSIL, respectively, and was 5% in hrHPV-positive but HPV 16/18-negative women with either ASC-US or LSIL. Methodological and technical heterogeneity among studies; insufficient data to assess accuracy of separate assays. Testing for HPV 16/18 to triage women with minor abnormal cytology is poorly sensitive but may be useful as a second triage test after hrHPV testing, with direct referral if the woman is positive for HPV 16/18. Whether colposcopy or repeated testing is recommended for hrHPV-positive but HPV 16/18-negative women depends on local decision thresholds that can be derived from pretest-posttest probability plots. 7th Framework Programme of the European Commission.

Triage of ASC-H: A meta-analysis of the accuracy of high-risk HPV testing and other markers to detect cervical precancer.

Women with a cytological diagnosis of atypical squamous cells, cannot exclude high-grade squamous intraepithelial lesion (ASC-H) are usually immediately referred for colposcopy. However, triage may reduce the burden of the diagnostic workup and prevent overtreatment. A meta-analysis was conducted to assess the accuracy of high-risk human papillomavirus (hrHPV) testing and testing for other molecular markers for the detection of grade 2 cervical intraepithelial neoplasia or worse (CIN2+) or grade 3 cervical intraepithelial neoplasia or worse (CIN3+) in women with ASC-H. An additional question that was assessed was whether triage would be useful in light of the relatively high pretriage probability of underlying precancer. The pooled absolute sensitivity and specificity of the Hybrid Capture 2 (HC2) assay for CIN2 + (derived from 19 studies) were 93% (95% confidence interval [CI], 89%-95%) and 45% (95% CI, 41%-50%), respectively. p16(INK4a) staining (only 3 studies) had similar sensitivity (93%; 95% CI, 75%-100%) but superior specificity (specificity ratio, 1.69) to HC2 for CIN2+. Testing for paired box 1 gene methylation (only 1 study) showed a superior specificity of 95% (specificity ratio, 2.08). The average pretest risk was 34% for CIN2 + and 20% for CIN3+. A negative HC2 result decreased this to 8% and 5%, respectively, whereas a positive result upgraded the risk to 47% and 28%, respectively. Because of the high probability of precancer with a diagnosis of ASC-H, the utility of triage is limited. The usual recommendation for referring women with ASC-H for colposcopy is not altered by a positive triage test, whatever test is used. A negative hrHPV DNA or p16(INK4a) test may allow repeat testing, but this recommendation will depend on local decision thresholds for referral.

Optimisation of cooperative spectrum sensing with ultra-wideband signalling

Reliable detection of primary user activity by spectrum sensing is of fundamental importance for cognitive radio systems. The objective of cooperative spectrum sensing is to combine the detection results of multiple cognitive radios in order to maximise the probability of detecting unused spectrum while meeting a required reliability of detecting primary user activity. In this paper, an optimisation strategy for cooperative spectrum sensing is proposed consisting of two parts. The first part is a Kullback-Leibler distance-based optimisation approach for the local decision thresholds. For the transmission of the local decisions, which can be realised by ultra-wideband underlay communications, in the second part an application-specific power control strategy is presented. The feasibility and effectiveness of the concepts are validated by real spectrum sensing results. The employed practical setup is based on software defined radio and detects a WiMAX-like OFDM signal.

Adaptive Cooperative Spectrum Sensing Using Group Intelligence

Opportunistic Spectrum Access in Cognitive Radios (CRs) calls for efficient and accurate spectrum sensing mechanism that provides the CR network with current spectral occupancy information. For a CR using energy detection for spectrum sensing, exact knowledge of Signal to Noise Ratio (SNR) at the receiver is crucial for determination of the decision threshold. This threshold in turn determines the probability of error (Probability of missed detection and probability of false alarm). In this paper, an innovative technique is proposed wherein spectral occupancy decisions from different CRs are combined and used as a training signal to adapt the local decision threshold. Each CR trains itself such that its decision is in alignment with other CRs in the network. Same can be looked at from group intelligence perspective where, multiple users, each with incomplete information, can learn from the group’s wisdom to reach a supposedly correct conclusion. Simulations under Rayleigh fading show probability of error at par with other co-operative spectrum sensing techniques albeit at lower complexity levels. We also probe into the accuracy of those decisions with standard techniques from a Cognitive Network perspective to prove the wisdom in group knowledge.

A flexible cooperative spectrum sensing scheme for cognitive radio networks

In this paper, we propose a three-threshold decision based cooperative spectrum sensing algorithm (TTDCSS) with two rounds cooperation. Specifically, after the failure of the first cooperation, cognitive users begin to report the second local decision bit to the fusion center (FC) to perform a second cooperation, thus sensing failure is eliminated. Furthermore, by integrating uncertainty of noise variance, we propose a simple method to set local decision thresholds. Numerical results show that the TTDCSS algorithm is superior to the conventional double-threshold decision based sensing methods in the performance with a little increase of sensing bits.

Diagnostic Accuracy of Serum Ceruloplasmin in Wilson Disease: Determination of Sensitivity and Specificity by ROC Curve Analysis among ATP7B-Genotyped Subjects

A serum ceruloplasmin concentration below 0.20 g/L is conventionally considered as one of the major diagnostic criteria for Wilson disease. This decision threshold has not been fully validated for its diagnostic characteristics, however. In this study, we evaluated various decision thresholds of serum ceruloplasmin concentration in the diagnosis of Wilson disease based on genotype-verified Wilson disease patients, carriers, and normal individuals. Serum ceruloplasmin concentration was measured by a nephelometric method in 57 Wilson disease patients and 71 family members (49 heterozygotes and 22 wild-type homozygotes), a validation group of 25 subjects clinically suspected of Wilson disease, and 690 normal individuals. We performed ROC analysis using Analyze-it software and confirmed the genotypes by direct DNA sequencing of ATP7B. Serum ceruloplasmin concentrations <0.20, 0.14, and 0.10 g/L showed positive predictive values of 48.3%, 100%, and 100%, respectively, and negative predictive values of 98.7%, 97.1%, and 91.9%. In the validation group, a serum ceruloplasmin threshold of 0.14 g/L rendered 100% sensitivity and specificity. Forty of 690 healthy subjects had serum ceruloplasmin concentrations <0.20 g/L; however, these 40 individuals had normal genotypes by DNA sequencing, and none of the 40 had ceruloplasmin concentrations <0.14 g/L. The diagnostic accuracy for Wilson disease using a serum ceruloplasmin concentration of 0.14 g/L as the local decision threshold was better than that using a threshold of 0.20 g/L. We suggest that laboratories providing ceruloplasmin assays determine decision thresholds based on local populations.

Distributed M-ary hypothesis testing with binary local decisions

Parallel distributed detection schemes for M-ary hypothesis testing often assume that for each observation the local detector transmits at least log2 M bits to a data fusion center (DFC). However, it is possible for less than log2 M bits to be available, and in this study we consider 1-bit local detectors with M > 2. We develop conditions for asymptotic detection of the correct hypothesis by the DFC, formulate the optimal decision rules for the DFC, and derive expressions for the performance of the system. Local detector design is demonstrated in examples, using genetic algorithm search for local decision thresholds. We also provide an intuitive geometric interpretation for the partitioning of the observations into decision regions. The interpretation is presented in terms of the joint probability of the local decisions and the hypotheses. 2003 Elsevier B.V. All rights reserved.