Improve Discrimination Ability Research Articles

DNA Barcoding of Invasive Terrestrial Plant Species in India.

Invasive plants are known to cause biodiversity loss and pose a major risk to human health and environment. Identification of invasive plants and distinguishing them from native species has been relied on morphological examination. Stringent requirement of floral characters and decreasing number of expert taxonomists are making conventional morphology-based identification system tedious and resource-intensive. DNA barcoding may help in quick identification of invasive species if distinct sequence divergence pattern at various taxonomic levels is observed. The present work evaluates the utility of four molecular markers; rbcL, matK, their combination (rbcL + matK), and psbA-trnH for identification of 37 invasive plant species from India and also in distinguishing them from 97 native species. A psbA-trnH locus was found to be of restricted utility in this work as it was represented by the members of a single family. A hierarchical increase in K2P mean divergence across different taxonomic levels was found to be the maximum for matK alone followed by rbcL + matK and rbcL alone, respectively. NJ clustering analysis, however, confirmed the suitability of combined locus (rbcL + matK) over individual rbcL and matK as the DNA barcode. RbcL showed the lowest resolution power among the three markers studied. MatK exhibited much better performance compared to rbcL alone in identifying most of the species accurately although it failed to show monophyly of genus Dinebra. Two families; Asteraceae and Poaceae, remained polyphyletic in the trees constructed by all three markers. Combined locus (rbcL + matK) was found to be the most suitable marker as it raised the resolution power of both the markers and could identify more than 90% of genera correctly. Phylogenetic tree constructed by Maximum-Parsimony method using combined locus as a molecular marker exhibited the best resolution, thus, supporting the significance of two-locus combination of rbcL + matK for barcoding invasive plant species from India. Present study contributes to the global barcode data of invasive plant species by adding fifty-one new sequences to it. Effective barcoding of additional number of native as well as invasive plant species from India is possible using this dual locus if it is combined with one or more new molecular plastid markers. Expansion of barcode database with a focus on barcode performance optimisation to improve discrimination ability at species level can be undertaken in future.

Chemical Nose Strategy with Metabolic Labeling and "Antibiotic-Responsive Spectrum" Enables Accurate and Rapid Pathogen Identification.

The worldwide antimicrobial resistance (AMR) dilemma urgently requires rapid and accurate pathogen phenotype discrimination and antibiotic resistance identification. The conventional protocols are either time-consuming or depend on expensive instrumentations. Herein, we demonstrate a metabolic-labeling-assisted chemical nose strategy for phenotyping classification and antibiotic resistance identification of pathogens based on the "antibiotic-responsive spectrum" of different pathogens. d-Amino acids with click handles were metabolically incorporated into the cell wall of pathogens for further clicking with dibenzocyclooctyne-functionalized upconversion nanoparticles (DBCO-UCNPs) in the presence/absence of six types of antibiotics, which generates seven-channel sensing responses. With the assistance of machine learning algorithms, eight types of pathogens, including three types of antibiotic-resistant bacteria, can be well classified and discriminated in terms of microbial taxonomies, Gram phenotypes, and antibiotic resistance. The present metabolic-labeling-assisted strategy exhibits good anti-interference capability and improved discrimination ability rooted in the unique sensing mechanism. Sensitive identification of pathogens with 100% accuracy from artificial urinary tract infection samples at a concentration as low as 105 CFU/mL was achieved. Pathogens outside of the training set can also be discriminated well. This clearly demonstrated the potential of the present strategy in the identification of unknown pathogens in clinical samples.

Video-based assessment of intraoperative surgical skill.

Surgeons' skill in the operating room is a major determinant of patient outcomes. Assessment of surgeons' skill is necessary to improve patient outcomes and quality of care through surgical training and coaching. Methods for video-based assessment of surgical skill can provide objective and efficient tools for surgeons. Our work introduces a new method based on attention mechanisms and provides a comprehensive comparative analysis of state-of-the-art methods for video-based assessment of surgical skill in the operating room. Using a dataset of 99 videos of capsulorhexis, a critical step in cataract surgery, we evaluated image feature-based methods and two deep learning methods to assess skill using RGB videos. In the first method, we predict instrument tips as keypoints and predict surgical skill using temporal convolutional neural networks. In the second method, we propose a frame-wise encoder (2D convolutional neural network) followed by a temporal model (recurrent neural network), both of which are augmented by visual attention mechanisms. We computed the area under the receiver operating characteristic curve (AUC), sensitivity, specificity, and predictive values through fivefold cross-validation. To classify a binary skill label (expert vs. novice), the range of AUC estimates was 0.49 (95% confidence interval; CI = 0.37 to 0.60) to 0.76 (95% CI = 0.66 to 0.85) for image feature-based methods. The sensitivity and specificity were consistently high for none of the methods. For the deep learning methods, the AUC was 0.79 (95% CI = 0.70 to 0.88) using keypoints alone, 0.78 (95% CI = 0.69 to 0.88) and 0.75 (95% CI = 0.65 to 0.85) with and without attention mechanisms, respectively. Deep learning methods are necessary for video-based assessment of surgical skill in the operating room. Attention mechanisms improved discrimination ability of the network. Our findings should be evaluated for external validity in other datasets.

Plasma Phosphorylated-tau181 Is a Predictor of Post-stroke Cognitive Impairment: A Longitudinal Study.

IntroductionPost-stroke cognitive impairment (PSCI) cannot be neglected because it drastically influences the daily life of patients and their families. However, there are no studies exploring the association between preclinical blood biomarkers of neurodegeneration including plasma amyloid-β (Aβ), tau, and brain-derived neurotrophic factor (BDNF) together with the risk of PSCI. This longitudinal study was to investigate whether these blood biomarkers with imaging markers of cerebral small vessel disease can improve the prediction for PSCI. In addition, we also explored the association between blood biomarkers with the trajectories of PSCI.MethodsAdult patients with first-ever acute ischemic stroke were recruited, and the cognitive and functional abilities of these patients were evaluated. Furthermore, blood biomarkers of neurodegeneration including plasma Aβ-40, Aβ-42, total tau, phosphorylated tau 181 (p-tau181), and BDNF levels and image markers of cerebral small vessel disease were measured. Each patient was followed up at 3 and 12 months at the outpatient department.ResultsOf 136 patients, 40 and 50 patients developed PSCI at 3 and 12 months after stroke, respectively. In functional trajectories, 27 patients did not have PSCI at 3 months but did at 12 months. By contrast, the PSCI status of 17 patients at 3 months was reversed at 12 months. Patients with high-acute plasma p-tau181 had a significantly lower PSCI risk at 3 months (odds ratio [OR] = 0.62, 95% CI = 0.40–0.94, p = 0.0243) and 12 months (OR = 0.69, 95% CI = 0.47–0.99, p = 0.0443) after adjustment for covariates and image biomarkers. Discrimination and reclassification statistics indicated that the p-tau181 level can improve discrimination ability for PSCI at 3 and 12 months, respectively. In addition, the plasma p-tau181 level was the highest in subjects without PSCI followed by those with delayed-onset PSCI and early-onset PSCI with reversal, whereas the lowest plasma p-tau181 level was found among those with persistent PSCI, showing a significant trend test (p = 0.0081).ConclusionPlasma p-tau181 is a potential biomarker for predicting early- and delayed-onset PSCI. Future studies should incorporate plasma p-tau181 as an indicator for timely cognitive intervention in the follow-up of patients with stroke.

Non-Laboratory-Based Simple Screening Model for Nonalcoholic Fatty Liver Disease in Patients with Type 2 Diabetes Developed Using Multi-Center Cohorts.

BackgroundNonalcoholic fatty liver disease (NAFLD) is the most prevalent cause of chronic liver disease worldwide. Type 2 diabetes mellitus (T2DM) is a risk factor that accelerates NAFLD progression, leading to fibrosis and cirrhosis. Thus, here we aimed to develop a simple model to predict the presence of NAFLD based on clinical parameters of patients with T2DM.MethodsA total of 698 patients with T2DM who visited five medical centers were included. NAFLD was evaluated using transient elastography. Univariate logistic regression analyses were performed to identify potential contributors to NAFLD, followed by multivariable logistic regression analyses to create the final prediction model for NAFLD.ResultsTwo NAFLD prediction models were developed, with and without serum biomarker use. The non-laboratory model comprised six variables: age, sex, waist circumference, body mass index (BMI), dyslipidemia, and smoking status. For a cutoff value of ≥60, the prediction accuracy was 0.780 (95% confidence interval [CI], 0.743 to 0.817). The second comprehensive model showed an improved discrimination ability of up to 0.815 (95% CI, 0.782 to 0.847) and comprised seven variables: age, sex, waist circumference, BMI, glycated hemoglobin, triglyceride, and alanine aminotransferase to aspartate aminotransferase ratio. Our non-laboratory model showed non-inferiority in the prediction of NAFLD versus previously established models, including serum parameters.ConclusionThe new models are simple and user-friendly screening methods that can identify individuals with T2DM who are at high-risk for NAFLD. Additional studies are warranted to validate these new models as useful predictive tools for NAFLD in clinical practice.

Prognostic value of the systemic immune-inflammation index in non-muscle invasive bladder cancer

PurposeWe assessed the prognostic value of systemic immune-inflammation index (SII) to refine risk stratification of the heterogeneous spectrum of patients with non-muscle-invasive bladder cancer (NMIBC)MethodsIn this multi-institutional cohort, preoperative blood-based SII was retrospectively assessed in 1117 patients with NMIBC who underwent transurethral resection of bladder (TURB) between 1996 and 2007. The optimal cut-off value of SII was determined as 580 using the best Youden index. Cox regression analyses were performed. The concordance index (C-index) and decision curve analysis (DCA) were used to assess the discrimination of the predictive models.ResultsOverall, 309 (28%) patients had high SII. On multivariable analyses, high SII was significantly associated with worse PFS (hazard ratio [HR] 1.84; 95% confidence interval [CI] 1.23–2.77; P = 0.003) and CSS (HR 2.53; 95% CI 1.42–4.48; P = 0.001). Subgroup analyses, according to the European Association of Urology guidelines, demonstrated the main prognostic impact of high SII, with regards to PFS (HR 3.39; 95%CI 1.57–7.31; P = 0.002) and CSS (HR 4.93; 95% CI 1.70–14.3; P = 0.005), in patients with intermediate-risk group; addition of SII to the standard predictive model improved its discrimination ability both on C-index (6% and 12%, respectively) and DCA. In exploratory intergroup analyses of patients with intermediate-risk, the improved discrimination ability was retained the prediction of PFS and CSS.ConclusionPreoperative SII seems to identify NMIBC patients who have a worse disease and prognosis. Such easily available and cheap standard biomarkers may help refine the decision-making process regarding adjuvant treatment in patients with intermediate-risk NMIBC.

Application of Unsupervised Learning Techniques to Identify Atlantic Tropical Cyclone Rapid Intensification Environments

AbstractTropical cyclone (TC) track forecasts have improved in recent decades while intensity forecasts, particularly predictions of rapid intensification (RI), continue to show low skill. Many statistical methods have shown promise in predicting RI using environmental fields, although these methods rely heavily upon supervised learning techniques such as classification. Advances in unsupervised learning techniques, particularly those that integrate nonlinearity into the class separation problem, can improve discrimination ability for difficult tasks such as RI prediction. This study quantifies separability between RI and non-RI environments for 2004–16 Atlantic Ocean TCs using an unsupervised learning method that blends principal component analysis with k-means cluster analysis. Input fields consisted of TC-centered 1° Global Forecast System analysis (GFSA) grids (170 different variables and isobaric levels) for 3605 TC samples and five domain sizes. Results are directly compared with separability offered by operational RI forecast predictors for eight RI definitions. The unsupervised learning procedure produced improved separability over operational predictors for all eight RI definitions, five of which showed statistically significant improvement. Composites from these best-separating GFSA fields highlighted the importance of mid- and upper-level relative humidity in identifying the onset of short-term RI, whereas long-term, higher-magnitude RI was generally associated with weaker absolute vorticity. Other useful predictors included optimal thermodynamic RI ingredients along the mean trajectory of the TC. The results suggest that the orientation of a more favorable thermodynamic environment relative to the TC and midlevel vorticity magnitudes could be useful predictors for RI.

Sex- and age-related differences in the predictive capability of circulating biomarkers: from the MONICA 10 cohort

Objectives The purpose of this study was to assess whether high-sensitivity C-reactive protein (hs-CRP), N-terminal pro-brain natriuretic peptide (NT-proBNP), and soluble urokinase plasminogen activator receptor (suPAR) differed in their ability to predict cardiovascular outcomes beyond traditional risk factors in younger and older men and women without known cardiovascular disease. Design. Prospective population-based cohort study of 1951 individuals from the MONItoring of trends and determinants in Cardiovascular disease (MONICA) study, examined 1993–1994. Participants were stratified into four groups based on sex and age. Subjects aged 41 or 51 years were classified as younger; those aged 61 or 71 years were classified as older. The principal endpoint was death from cardiovascular causes. Predictive capabilities of biomarkers were tested using Cox proportional-hazards regression, Harrell’s concordance-index, net reclassification improvement, and classification and regression tree (CART) analysis. Results. Median follow-up was 18.5 years, during which 19/597 younger men, 100/380 older men, 12/607 younger women, and 46/367 older women had died from a cardiovascular cause. NT-proBNP was independently associated with death from cardiovascular causes among all participants (p ≤ .02) except younger women (p = .70), whereas hs-CRP was associated with this endpoint in men (p ≤ .007), and suPAR in older men only (p < .001). None of the biomarkers improved discrimination ability beyond traditional risk factors (p ≥ .07). However, NT-proBNP enhanced reclassification in men and older women. CART-analysis showed that NT-proBNP was generally of greater value among men, and suPAR among women. Conclusions. Hs-CRP, NT-proBNP, and suPAR displayed different associations with cardiovascular death among apparently healthy younger and older men and women.

Neural bases of visual processing of moving and stationary stimuli presented to the blind hemifield of hemianopic patients

Unilateral damage to post-chiasmatic visual pathways or cortical areas results in the loss of vision in the contralateral hemifield, known as hemianopia. Some patients, however, may retain the ability to perform an above chance unconscious detection or discrimination of visual stimuli presented to the blind hemifield, known as “blindsight”. An important finding in blindsight research is that it can often be elicited by moving stimuli. Therefore, in the present study, we wanted to test whether moving stimuli might yield blindsight phenomena in patients with cortical lesions resulting in hemianopia, in a discrimination task where stimulus movement is orthogonal to the feature of interest. This could represent an important strategy for rehabilitation because it might improve discrimination ability of stimulus features different but related to movement, e.g. line orientation.We tested eight hemianopic patients and eight age-matched healthy controls in an orientation discrimination task with moving or static visual stimuli. During performance of the task we carried out fMRI scanning and tractography. Behaviourally, we did not find a reliable main effect of motion on orientation discrimination; however, an important result was that in different patients blindsight could occur only with moving or stationary stimuli or with both. As to brain imaging results, following presentation of moving stimuli to the blind hemifield, a widespread fronto-parietal bilateral network was recruited including areas of the dorsal stream and in particular bilateral motion area hMT + whose activation positively correlated with behavioural performance. This bilateral network was not activated in controls suggesting that it represents a compensatory functional change following brain damage. Moreover, there was a higher activation of ipsilesional area hMT+ in patients who performed above chance in the moving condition. By contrast, in patients who performed above chance in the static condition, we found a higher activation of contralesional area V1 and extrastriate visual areas. Finally, we found a linear relationship between structural integrity of the ipsilesional pathway connecting lateral geniculate nucleus (LGN) with motion area hMT+ and both behavioural performance and ipsilesional hMT + activation. These results support the role of LGN in modulating performance as well as BOLD amplitude in the absence of visual awareness in ipsilesional area hMT+ during an orientation discrimination task with moving stimuli.

Effect of fundus tracking on structure–function relationship in glaucoma

AimsTo investigate the effect of fundus tracking perimetry on structure–function relationship in glaucoma.MethodsPerimetric data were acquired with the Humphrey Field Analyzer (HFA) and the Compass fundus perimeter (CMP, equipped with...

CT-pathologic correlation of non-calcified atherosclerotic arterial plaques: a study using carotid endarterectomy specimens

Pathologic features of atherosclerotic plaques on CT are not established. We compared CT values among pathologically confirmed plaque constituents and evaluated their ability to distinguish plaque constituents. 50 histopathological images of carotid endarterectomy samples from 10 males and 2 females (age 54-74 years, average 65.9 years) were examined. We compared pre-operative CT [pre-contrast (CT-P), early post-contrast phase (CT-E), delayed post-contrast phase (CT-D)] of lipid-rich necrotic core (NC) and fibrous tissue (F) plaque components with pathological images. The ability of features to differentiate plaque components using several discrimination techniques were compared. CT values of NC and F were 36 ± 13, 45 ± 11 (mean ± standard deviation, Hounsfield unit, HU), 41 ± 17, 69 ± 18, and 44 ± 16, 70 ± 13 in CT-P (p < 0.01), CT-E (p < 0.0001), and CT-D (p < 0.0001), respectively. The threshold, sensitivity, and accuracy for distinguishing NC from F were 44 HU, 74%, and 68%; 55 HU, 85%, and 85%; and 63 HU, 92%, and 84% in CTP, CT-E, and CT-D, respectively. CT-P had lower accuracy than CT-E and CT-D (both p < 0.05), but CT-E and CT-D were similar. CT-E and CT-D yielded 90 and 91% sensitivity and accuracy, respectively in linear discrimination analysis. In both pre- and post-contrast CT, CT values were lower in NC than F. Although values overlapped, using two-phase post-contrast CTs improved discrimination ability. Our findings may help to establish computer-aided diagnosis of vulnerable atherosclerotic plaques in future.

VISSA-PLS-DA-Based Metabolomics Reveals a Multitargeted Mechanism of Traditional Chinese Medicine for Traumatic Brain Injury.

Metabolomics is an emerging tool to uncover the complex pathogenesis of disease, as well as the multitargets of traditional Chinese medicines, with chemometric analysis being a key step. However, conventional algorithms are not suitable for directly analyzing data at all times. The variable iterative space shrinkage approach-partial least squares-discriminant analysis, a novel algorithm for data mining, was first explored to screen metabolic varieties to reveal the multitargets of Xuefu Zhuyu decoction (XFZY) against traumatic brain injury (TBI) by the 7th day. Rat plasma from Sham, Vehicle, and XFZY groups was used for gas chromatography/mass spectrometry-based metabolomics. This method showed an improved discrimination ability (area under the curve = 93.64%). Threonine, trans-4-hydroxyproline, and creatinine were identified as the direct metabolic targets of XFZY against TBI. Five metabolic pathways affected by XFZY in TBI rats, were enriched using Metabolic Pathway Analysis web tool (i.e., phenylalanine, tyrosine, and tryptophan biosynthesis; phenylalanine metabolism; galactose metabolism; alanine, aspartate, and glutamate metabolism; and tryptophan metabolism). In conclusion, metabolomics coupled with variable iterative space shrinkage approach-partial least squares-discriminant analysis model may be a valuable tool for identifying the holistic molecular mechanisms involved in the effects of traditional Chinese medicine, such as XFZY.

ROC analysis using covariate balancing propensity scores with an application to biochemical predictors for thyroid cancer

Biomarker evaluation is important for diagnosing clinical diseases. Covariate adjusted receiver operating characteristic (ROC) regression has been used to identify significant biomarker candidates. Here, we show that the statistical significance of a biomarker can be affected by its prevalence. We propose a novel method that incorporates covariate prevalence information in the ROC regression. This approach is based on covariate balancing propensity scores proposed by Imai and van Dyk. Our method produces higher AUC values, demonstrating improved discrimination ability compared to direct ROC regression or unadjusted ROC analysis; this method can be used to improve biomarker development and can be implemented by an artificial intelligence (AI) system. Extensive simulation studies and data from a thyroid cancer study illustrate the advantages of our approach.

A colorimetric sensor array for protein discrimination based on carbon nanodots-induced reversible aggregation of AuNP with GSH as a regulator

We have proposed a doubled nanoplasmonic colorimetric sensor array for protein recognition based on car-bon nanodots (CDs)-induced reversible aggregation of gold nanoparticle (AuNP) with GSH as a regulator. It is found that different kinds of CDs can induce the aggregation of AuNP with different degrees and cause different color change. Adding reduced glutathione (GSH) into the CDs/AuNP system can make the AuNP disperse again accompanying with a second color change. Based on this phenomenon, three kinds of CDs originated from different amino acids and citric acid are used to fabricate a colorimetric sensor array with three sensing elements. Upon addition of target proteins into CDs/AuNP conjugates, the competitive protein-AuNP affinity disrupts the SPR absorption peak of CDs/AuNP, generating cross-reactive three-unit colorimetric signal. Subsequently, GSH is introduced into each unit containing protein samples in situ. The SPR absorption peak of AuNP is substantially affected due to the stronger affinity between AuNP and GSH, producing an-other three-unit response signal. With this strategy, each sensing element can give two output signals to each target protein, which doubles the number of unit from three to six without changing the three-cell configuration, resulting in greatly improved discrimination ability of proteins and identification accuracy of unknown samples. Twelve proteins at 50 nM can be well discriminated by this novel colorimetric sensor array. Moreover, our sensing system is very stable in the high concen-tration of salt, make it suitable for the real sample analysis. This method effectively identifies twelve proteins at 200 nM spiked in human urine.

Predictive Modeling of Hospital Mortality for Patients With Heart Failure by Using an Improved Random Survival Forest

Identification of different risk factors and early prediction of mortality for patients with heart failure are crucial for guiding clinical decision-making in Intensive care unit cohorts. In this paper, we developed a comprehensive risk model for predicting heart failure mortality with a high level of accuracy using an improved random survival forest (iRSF). Utilizing a novel split rule and stopping criterion, the proposed iRSF was able to identify more accurate predictors to separate survivors and nonsurvivors and thus improve discrimination ability. Based on the public MIMIC II clinical database with 8059 patients, 32 risk factors, including demographics, clinical, laboratory information, and medications, were analyzed and used to develop the risk model for patients with heart failure. Compared with previous studies, more critical laboratory predictors were identified that could reveal difficult-to-manage comorbidities, including aspartate aminotransferase, alanine aminotransferase, total bilirubin, serum creatine, blood urea nitrogen, and their inherent effects on events; these were determined to be critical indicators for predicting heart failure mortality with the proposed iRSF. The experimental results showed that the developed risk model was superior to those used in previous studies and the conventional random survival forest-based model with an out-of-bag C-statistic value of 0.821. Therefore, the developed iRSF-based risk model could serve as a valuable tool for clinicians in heart failure mortality prediction.

Continuously evolving ‘chemical tongue’ biosensor for detecting proteins

We developed a unique continuously evolving colorimetric sensor array based on AuNPs decorated by two single-stranded oligonucleotides with different molar ratios for protein discrimination. The number of differential receptors in this sensor array could be easily extended by adjusting the molar ratios of two DNA, resulting in continuously improved discrimination ability. The continuous response data of target samples against our sensing system could be easily obtained and exclude abnormal signals. The sensing system could discriminate twelve proteins at the concentration of 200nM in the presence of 50% human urine with accuracy of 100%, showing feasible potential for diagnostic applications. Remarkably, HSA at various concentrations, the pure Lys and HSA, and the mixture of these two proteins with different molar ratios had been successfully discriminated in LDA plot as well in the presence of human urine sample. This novel strategy will be very promising for the design of cheaper and more reliable sensor arrays for target samples.

Degraded neural and behavioral processing of speech sounds in a rat model of Rett syndrome

Individuals with Rett syndrome have greatly impaired speech and language abilities. Auditory brainstem responses to sounds are normal, but cortical responses are highly abnormal. In this study, we used the novel rat Mecp2 knockout model of Rett syndrome to document the neural and behavioral processing of speech sounds. We hypothesized that both speech discrimination ability and the neural response to speech sounds would be impaired in Mecp2 rats. We expected that extensive speech training would improve speech discrimination ability and the cortical response to speech sounds. Our results reveal that speech responses across all four auditory cortex fields of Mecp2 rats were hyperexcitable, responded slower, and were less able to follow rapidly presented sounds. While Mecp2 rats could accurately perform consonant and vowel discrimination tasks in quiet, they were significantly impaired at speech sound discrimination in background noise. Extensive speech training improved discrimination ability. Training shifted cortical responses in both Mecp2 and control rats to favor the onset of speech sounds. While training increased the response to low frequency sounds in control rats, the opposite occurred in Mecp2 rats. Although neural coding and plasticity are abnormal in the rat model of Rett syndrome, extensive therapy appears to be effective. These findings may help to explain some aspects of communication deficits in Rett syndrome and suggest that extensive rehabilitation therapy might prove beneficial.

Identification of the geographical origins of pomelos using multielement fingerprinting.

Eighty pomelo samples and 80 soil samples were examined using a multielement component test to predict the geographical origins of pomelos produced in 4 regions (Sichuan, Chongqing, Fujian, and Guangxi Provinces) of China. The concentrations of 8 elements were determined by atomic absorption spectrometry. Ca, K, and Na were the most abundant elements. Principal component analysis (PCA) and linear discriminant analysis (LDA) were applied to reduce the dimensionality of the multielement data from 8 to 2 while retaining the highest possible variance. Using PCA and LDA, 69.66% and 91.30%, respectively, of the pomelo origins were classified correctly using multielement variables, along with 67.06% and 83.40% for soil multielement analysis. Results indicated that the LDA method was more effective for geographical origin classification than PCA. The results of the multielement component test demonstrated its capability to screen pomelo origins rapidly.

Discrimination of similar characters using nonlinear normalization based on regional importance measure

Discrimination of confusing characters is very important in recognition of character sets containing a multitude of similar characters. Confusing characters have very similar shapes and are separated by only a small difference. For a successful discrimination, we need to focus on that difference. However, the small difference can be reduced or even lost during the feature extraction process. In such a case, further analysis after the feature extraction rarely succeeds. This paper proposes a discriminative nonlinear normalization algorithm to improve discrimination ability. The proposed method emphasizes the difference between confusing characters. It measures the importance of each region in the discrimination of confusing characters. Then, it resamples the image according to the regional importance measure. As a result, it expands important regions but shrinks less important regions. Since it emphasizes important regions in the preprocessing step, it does not suffer from the information loss during the feature extraction. In experiments, the proposed method successfully detected and expanded important regions. In handwritten Hangul recognition, the proposed method outperformed other two recently developed pair-wise discrimination methods. On SERI95a data set, it improved the recognition rate from 87.69 to 90.11 %, achieving a 19.66 % error reduction rate.

SAR TARGET CLASSIFICATION USING BAYESIAN COMPRESSIVE SENSING WITH SCATTERING CENTERS FEATURES

The emerging fleld of compressed sensing provides sparse reconstruction, which has demonstrated promising results in the areas of signal processing and pattern recognition. In this paper, a new approach for synthetic aperture radar (SAR) target classiflcation is proposed based on Bayesian compressive sensing (BCS) with scattering centers features. Scattering centers features is extracted as a l1-norm sparse problem on the basis of SAR observation physical model, which can improve discrimination ability compared with original SAR image. Using an overcomplete dictionary constructed by training samples, BCS is utilized to design targets classifler. For target classiflcation performance evaluation, the proposed method is compared with several state-of-art methods through experiments on Moving and Stationary Target Acquisition and Recognition (MSTAR) public release database. Experiment results illustrate the efiectiveness and robustness of the proposed approach.