Discriminant Criterion Research Articles

Using Multiple Discriminant Analysis for the Assignment of Initial Water Entitlements at River Basin-Level under the Strictest Water Resources Management System Constraints in China

From the perspective of system science, in China, an assignment system of initial water entitlements at the river basin-level can be divided into two subsystems, namely the assignment subsystem of initial water entitlements at the province-level and government reserved water at the river basin-level. Under the new backdrop of implementing the strictest water resources management system (SWRMS), we propose a novel methodological framework for addressing the in-coordination of pre-allocation plans between two subsystems for Lake Tai Basin, China. First, considering total water use, pollutant discharge and water use efficiency, we establish several criteria for the discriminant analysis of pre-allocation plans. Whilst based on these criteria, we built a comprehensive discriminant criterion to further verify coupling and coordination of pre-allocation plans between two subsystems. Second, according to uncoordinated or less coordinated situations, we propose adjusted strategies to decide the direction (increase or decrease) of the adjustment for pre-allocation plans of two subsystems. Third, taking coupling and coordination as optimal objectives, and considering total water use, total pollutant discharge and water use efficiency as constraints, we built an adjusted decision-making model for the assignment of initial water entitlements of the basin. Finally, the results of this novel discriminant analysis methodology that were applied to the Lake Tai Basin show that under the water frequency of 75%, in the planning year 2030, Jiangsu Province is assigned the most initial water entitlements at the province-level, followed by Shanghai and Zhejiang Province. In this paper, results are generally in accordance with pilot plans released by China’s Ministry of Water Resources. Apart from pilot plans, our findings also show the assignment plans for government reserved water at the river basin-level, which is coordinated and coupled with assignment plans for initial water entitlements at the province-level. The novel methodological framework of this paper can also be applied as a reference to other similar river basin.

Multilinear Enhanced Fisher Discriminant Analysis for robust multimodal 2D and 3D face verification

In the last few years, multilinear subspace learning has attracted large interest for dimensionality reduction and classification of multidimentional data. In this paper, we propose a new multilinear supervised projection method, called Multilinear Enhanced Fisher Discriminant Analysis (MEFDA), which improves the generalization capacity by decomposing the discriminate analysis process into a simultaneous diagonalization of the inter-class and intra-class scatter matrices resulting from the unfolding of tensor data. MEFDA uses the discriminant tensor criterion based on iterative procedure in order to obtain the optimal projection matrix in each k mode of tensor data. One virtue of MEFDA is that it can avoid the disadvantage of large dimensionality, in which the computational cost is reduced. Since the data encoded using high-order tensor, MEFDA respects the original structure and the natural geometry of face data. Therefore, the natural correlation of multi-dimension face data can be exploited. The proposed algorithm has been evaluated on three public 3D face datasets, FRGC V2.0, CASIA 3D and Bosphorus under the hard challenges of poses, expressions, illuminations and occlusions. The experimental results demonstrate that the proposed method significantly outperforms the current state of the art methods, with 99% and 97.02% at 0.001 False Acceptance Rate (FAR) for FRGC v2.0, and Bosphorus datasets, respectively, and 99.36% at evaluation Equal Error Rate (EER) for CASIA 3D dataset.

L1 -Norm Heteroscedastic Discriminant Analysis Under Mixture of Gaussian Distributions.

Fisher's criterion is one of the most popular discriminant criteria for feature extraction. It is defined as the generalized Rayleigh quotient of the between-class scatter distance to the within-class scatter distance. Consequently, Fisher's criterion does not take advantage of the discriminant information in the class covariance differences, and hence, its discriminant ability largely depends on the class mean differences. If the class mean distances are relatively large compared with the within-class scatter distance, Fisher's criterion-based discriminant analysis methods may achieve a good discriminant performance. Otherwise, it may not deliver good results. Moreover, we observe that the between-class distance of Fisher's criterion is based on the l2 -norm, which would be disadvantageous to separate the classes with smaller class mean distances. To overcome the drawback of Fisher's criterion, in this paper, we first derive a new discriminant criterion, expressed as a mixture of absolute generalized Rayleigh quotients, based on a Bayes error upper bound estimation, where mixture of Gaussians is adopted to approximate the real distribution of data samples. Then, the criterion is further modified by replacing l2 -norm with l1 one to better describe the between-class scatter distance, such that it would be more effective to separate the different classes. Moreover, we propose a novel l1 -norm heteroscedastic discriminant analysis method based on the new discriminant analysis (L1-HDA/GM) for heteroscedastic feature extraction, in which the optimization problem of L1-HDA/GM can be efficiently solved by using the eigenvalue decomposition approach. Finally, we conduct extensive experiments on four real data sets and demonstrate that the proposed method achieves much competitive results compared with the state-of-the-art methods.

Geochemistry and Biomarker Analysis of the Bentonites from Esquivias (Toledo, Spain)

The Tajo Basin is one of the richest in Mg-clays known around the world. Mg-bentonites, kwon as pink clays and green clays, alter in the Intermediate Unit of the Miocene sediments. In this work, a new approach to the genesis of these bentonites is performed by studying for the first time the biomarkers present in these clays in relation to the mineralogy and geochemistry, as well as using discriminatory criteria between green and pink clays. Samples were collected at a quarry of Mg-bentonites, in the proximities of Esquivias (Toledo, Spain). Mineralogical characterization and semi-quantification (X-ray diffraction (XRD)) allowed a well-defined classification according to the mineralogical content of the samples to be established, differentiating four associations. Geochemical analyses are clearly linked to the mineralogy and provide criteria to differentiate the genesis of the materials studied. In this regard, green clays are interpreted as having a more detrital character than pink clays, which present a more authigenic character. Biomarkers (n-alkanes and n-alkanoic acids) were studied, not showing a clear link with the mineralogy as in the case of the geochemistry, but providing interesting information about the origin and degradation of the organic matter. Pink clays have higher contents in biomarkers than green clays, providing a discriminative criterion.

Respiratory Muscle Strength as a Discriminator of Sarcopenia in Community-Dwelling Elderly: A Cross-Sectional Study.

To compare the values obtained from maximum respiratory pressures (MRP) between sarcopenic and non-sarcopenic elderly; to verify the association of maximum respiratory pressures with sarcopenia and its indicators; and to establish cut-off points for MRP as a discriminator of sarcopenia. Cross-sectional study. Macapá, Brazil. Community-dwelling elderly ≥ 60 years old, both sexes. Evaluation of respiratory muscle strength (maximal inspiratory pressure - MIP and maximal expiratory pressure - MEP) and sarcopenia, according to the European Working Group on Sarcopenia in Older People (EWGSOP), in which the diagnosis of this condition considered the reduction of muscle mass (muscle mass index - MMI) associated with muscle strength reduction (hand grip strength - HGS) and / or impairment in physical performance (gait speed - GS). The sample consisted of 383 elderly individuals, with a mean age of 70.02 ± 7.3 years and a prevalence of sarcopenia of 12.53% (n = 48). Sarcopenic individuals presented significantly lower (obtained, obtained versus predicted) mean values for the maximal respiratory pressures compared to the non-sarcopenic elderly, and these were inversely associated with sarcopenia (an increase by 1 cmH2O in MIP and MEP reduced by 5% and 3%, respectively, the probability of sarcopenia). In relation to the association with the sarcopenia indicators, the increase by 1 cmH2O in MIP and MEP decreased, respectively, the probability of decreasing muscle strength (3% and 2%), GS (3% and 4%) and MMI (3 % - MIP). Cut-off points ≤60 cmH2O and ≤50 cmH2O for MEP and ≤55 cmH2O and ≤45 cmH2O for MEP, respectively for elderly men and women, served as a discriminant criterion for the presence of sarcopenia (area under the ROC curve superior to 0.70). Elderly patients with sarcopenia had lower MIP and MEP values when compared to non-sarcopenic individuals, and respiratory muscle strength was inversely associated with the diagnosis of sarcopenia and its indicators (HGS, gait speed and MMI). Furthermore, cut-off points for MIP and MEP can be used in clinical practice as discriminators of sarcopenia in community-dwelling elderly.

Time Spent in Sedentary Behaviour as Discriminant Criterion for Frailty in Older Adults.

This paper aims to analyse whether time spent in sedentary behaviour was a discriminant criterion for frailty in older adults. This was a cross-sectional study conducted in a sample of 457 elderly individuals aged ≥60 years. Frailty was defined as the presence of three or more of the following criteria: Unintentional weight loss, low walking speed at a 4.57 m course, reduced manual grip strength, exhaustion and insufficient physical activity level. Participants were classified into two groups: Non-frail or frail. Exposure to sedentary behaviour was assessed by the time spent sitting during a typical week, according to the adapted version of the International Physical Activity Questionnaire. Descriptive (mean, frequency) and inferential statistics (Poisson regression, Pearson’s Chi-square, Receiver Operating Characteristic Curve) were used to analyse the data, comparing them to the time-related areas exposed to sedentary behaviour by gender and the presence of fragility. The prevalence of frailty was 22.1% (n = 101). The most accurate cut-off points of sitting time for predicting frailty were >495 min/day (men) or >536 min/day (women). Time spent in sedentary behaviour can be used to indicate fragility in the elderly of both sexes.

Discriminant ability and criterion validity of the Trunk Impairment Scale for cerebral palsy

Aims: To compare the performance of children with mild and moderate-to-severe cerebral palsy (CP) on the Trunk Impairment Scale (TIS), Gross Motor Function Measure (GMFM), and on center-of-pressure variables; to establish the discriminant ability of these tools to predict severity of motor impairment in CP; and to investigate the criterion validity of the TIS.Methods: Children with mild (n = 18, 11 males, 7 females, mean age = 9.5 ± 2.9 years, Gross Motor Function Classification System I–II) and moderate-to-severe (n = 18, 11 males, 7 females, mean age = 9.2 ± 229, Gross Motor Function Classification System III–IV) CP were tested using the TIS and the GMFM, and during static sitting on force-plate.Results: Children with mild CP showed better trunk (median; 95% confidence interval = 22.5; 21.29–22.59 vs. 13; 11.97–14.8; p < 0.001) and gross motor (60; 57.73–59.3 vs. 40; 38.96–46.25; p < 0.001) scores, and better postural control (lower center of pressure (CoP) displacement [anterior–posterior: (0.42; 0.32–1.11 vs. 0.89; 0.70–1.65; p = 0.022); medial–lateral: (0.42; 0.31–1.08 vs. 0.91; 0.65–1.17; p = 0.044)], and lower area of sway, (0.05; −0.15–0.97 vs. 0.44; 0.23–0.90; p = 0.008) than the moderate-to-severe group. Trunk control and gross motor function explained 81.5% of the variance in the severity of motor condition. Correlations between the TIS and the GMFM were excellent (ρ = 0.944, p < 0.001); correlations between the TIS and CoP variables were low (anterior–posterior displacement: ρ = −0.411, p < 0.05; medial–lateral displacement: ρ = −0.327, p < 0.05); area of sway: ρ = −0.430, p < 0.05; velocity of sway: ρ = −0.308, p < 0.05).Conclusions: The TIS is able to differentiate levels of trunk control across various levels of motor impairments in CP. It is a valid tool to assess trunk control, showing very high concurrent validity with the GMFM sitting dimension.Implications for RehabilitationTrunk Impairment Scale (TIS) can be used by rehabilitation professionals to differentiate levels of trunk control across levels of motor impairment.TIS showed concurrent validity with Gross Motor Function Measure and should be used to assess trunk control in children with cerebral palsy (CP) in clinical settings.The use of TIS allows a reliable assessment of postural control in children with CP in clinical settings.

Protein subnuclear localization based on a new effective representation and intelligent kernel linear discriminant analysis by dichotomous greedy genetic algorithm.

A wide variety of methods have been proposed in protein subnuclear localization to improve the prediction accuracy. However, one important trend of these means is to treat fusion representation by fusing multiple feature representations, of which, the fusion process takes a lot of time. In view of this, this paper novelly proposed a method by combining a new single feature representation and a new algorithm to obtain good recognition rate. Specifically, based on the position-specific scoring matrix (PSSM), we proposed a new expression, correlation position-specific scoring matrix (CoPSSM) as the protein feature representation. Based on the classic nonlinear dimension reduction algorithm, kernel linear discriminant analysis (KLDA), we added a new discriminant criterion and proposed a dichotomous greedy genetic algorithm (DGGA) to intelligently select its kernel bandwidth parameter. Two public datasets with Jackknife test and KNN classifier were used for the numerical experiments. The results showed that the overall success rate (OSR) with single representation CoPSSM is larger than that with many relevant representations. The OSR of the proposed method can reach as high as 87.444% and 90.3361% for these two datasets, respectively, outperforming many current methods. To show the generalization of the proposed algorithm, two extra standard datasets of protein subcellular were chosen to conduct the expending experiment, and the prediction accuracy by Jackknife test and Independent test is still considerable.

A Contactless Rotation-Invariant Palm Vein Recognition System

There is a high demand for contactless biometrics due to social concern for more convenient and hygienic technology. However, contactless biometrics like palm vein recognition faces a number of challenges. Hand rotation is among the main challenges to be solved. A number of methods have been introduced to extract palm region of interest (ROI) and have reported high accuracy, but the maximum rotation that can be handled by the methods are never mentioned. Usually, the ROI detection method fails to work if a large rotation exists in the hand position. In this paper, we propose a rotation-invariant method that can detect a palm ROI in any rotation angles. The method can sustain the recognition accuracy despite a substantial hand rotation of up to 360°. ROI detection in different angles causes minor ROI dislocation, thus a feature extraction technique called Winner-Take-All hashing (WTA) is developed. WTA serves as a feature enhancer, and is equipped with the Fisher’s Discriminant criterion to obtain a discriminative representation from the palm ROI. Experiment result shows that the proposed method is able to achieve a maximum recall of 94% and an EER at 7.12% for a challenging scenario containing hand rotation of up to 360°.

Singular approach to penetrometry by preprocessing of digitized force-displacement curves and chemometry: A case study of 12 tomato varieties.

The force-displacement curve of penetrometry can be considered a "mechanical imprint" to improve fruit classification according to preharvest and postharvest characteristics (i.e., maturity, variety, etc.). This mechanical imprint represents a new and discriminant phenotyping criterion for tomato breeding to improve texture. The use of the digitized curve rather than computed parameters may be more easily implemented in industry.

New sunspots and aurorae in the historical Chinese text corpus? Comments on uncritical digital search applications

We review some applications of the method of electronic searching for historical observations of sunspots and aurorae in the Chinese text corpus by Hayakawa et al. (, , , ), Kawamura et al. (), and Tamazawa et al. (). However, we show strong shortcomings in the digital search technique as applied by them: almost all likely true sunspot and aurora records were presented earlier (e.g., Xu et al. ), which are not mentioned in those papers; the remaining records are dubious and often refer to other phenomena, neither spots nor aurorae (this also applies to Hayakawa et al. ). The alleged aurorae in Hayakawa et al. () and Kawamura et al. () show a broad peak around full moon, not expected for aurorae. Hayakawa et al. () use the Korean report At night, the gate of heaven was opened (between AD 992 Dec 6 and AD 993 Jan 25, i.e., close to the 14C variation AD 993/994) to estimate the Dst index of solar activity, even though the text does not fulfill any discriminative aurora criteria (except night‐time). Most of the above publications include very few Chinese texts and translations, and their tables with abbreviated keywords do not allow the reader to consider alternative interpretations (the tables also do not specify which records mention night‐time). We have compared some of their event tables with previously published catalogues and found various discrepancies. There are also intrinsic inconsistencies, misleading information (lunar phase for day‐time events), and dating errors. We present Chinese texts and translations for some of their presumable new aurorae: only one can be considered a likely true aurora (AD 604 Jan); some others were selected on the sole basis of the use of the word light or rainbow. Several alleged new aurorae present observations beside the Sun during day‐time. There are well‐known comets among their presumable aurorae. We also discuss (a) whether heiqi ri pang can stand for black spot(s) on one side of or beside the sun, (b) the aurora color confusion in Hayakawa et al. (, ), and (c) whether white and unusual rainbows can be aurorae.

Progressive Joint Modeling in Unsupervised Single-Channel Overlapped Speech Recognition

Unsupervised single-channel overlapped speech recognition is one of the hardest problems in automatic speech recognition (ASR). Permutation invariant training (PIT) is a state of the art model-based approach, which applies a single neural network to solve this single-input, multiple-output modeling problem. We propose to advance the current state of the art by imposing a modular structure on the neural network, applying a progressive pretraining regimen, and improving the objective function with transfer learning and a discriminative training criterion. The modular structure splits the problem into three sub-tasks: frame-wise interpreting, utterance-level speaker tracing, and speech recognition. The pretraining regimen uses these modules to solve progressively harder tasks. Transfer learning leverages parallel clean speech to improve the training targets for the network. Our discriminative training formulation is a modification of standard formulations, that also penalizes competing outputs of the system. Experiments are conducted on the artificial overlapped Switchboard and hub5e-swb dataset. The proposed framework achieves over 30% relative improvement of WER over both a strong jointly trained system, PIT for ASR, and a separately optimized system, PIT for speech separation with clean speech ASR model. The improvement comes from better model generalization, training efficiency and the sequence level linguistic knowledge integration.

Learning to rank in PRISM

Learning parameters associated with propositions is one of the main tasks of probabilistic logic programming (PLP), and learning algorithms for PLP have been primarily developed based on maximum likelihood estimation or the optimization of discriminative criteria. This paper explores yet another innovative approach to parameter learning, learning to rank or rank learning, that has been studied mainly in the field of preference learning. We combine learning to rank with techniques developed in PLP to make the latter applicable to a variety of ranking problems such as information retrieval. We implement our approach in PRISM, a PLP system based on the distribution semantics. It supports many parameter learning algorithms such as the expectation maximization algorithm, the variational Bayes algorithm and an algorithm for Viterbi training efficiently by mapping them onto a single data structure called explanation graph. To ensure the same efficiency for parameter learning by learning to rank as in the current PRISM, we introduce a gradient-based learning method that takes advantage of dynamic programming on the explanation graph. This paper also presents three experimental results. The first one is with synthetic data to check the learning behaviors of the proposed approach. The second one uses a knowledge base (knowledge graph) and apply rank learning to a DistMult model for the task of deciding whether relations over entities exist or not. The last one tackles the problem of parsing by a probabilistic context free grammar whose parameters are learned from a tree corpus by rank learning. These experiments successfully demonstrated the potential and effectiveness of learning to rank in PLP. We plan to release a new version of PRISM augmented with the ability of learning to rank in the near future.

Comparison of psychometric properties of subjective structured assessment instruments of technical performance during knee arthroscopy

ObjectivesVarious subjective assessment global rating scales to assess proficiency of surgeons while performing knee arthroscopy have been described but it is still unclear as to which of the assessment scale is the most optimal and “gold standard”. The aim of the present study was to compare and contrast psychometric properties like validity and reliability of objective assessment global rating scales for knee arthroscopy. MethodsA systematic review was performed of articles published in Pubmed, Embase, AMED, ERIC, Proquest and CINAHL. Ten assessment scales (Arthroscopic surgery skill evaluation tool [ASSET]; Arthroscopy skills score; global ratings of arthroscopic performance; basic arthroscopic knee scoring system [BAKSS]; modified basic arthroscopic knee scoring system [modified BAKSS]; modified objective structured assessment of technical skill [modified OSATS]; modified objective assessment of arthroscopic skills [modified OAAS]); modified orthopaedic competency assessment project [modified OCAP; Imperial global arthroscopy rating scale [IGARS] and Injury grading index [IGI] fulfilled the inclusion criteria. A predefined checklist was used to compare content validity, construct - convergent validity, construct - discriminant validity, criterion validity, predictive validity, internal consistency, inter-rater reliability, intra-rater reliability and test-retest reliability of all the included assessment global rating scales. ResultsThe ASSET demonstrated optimal number of satisfactory psychometric properties of content validity, construct-convergent validity, construct - discriminant validity, criterion validity, internal consistency, inter-rater reliability, intra-rater reliability and test-retest reliability. The IGI demonstrated the least number of satisfactory psychometric properties. ConclusionWe recommend the use of Arthroscopic surgery skill evaluation tool [ASSET] as it demonstrates optimal psychometric properties. The ASSET can be used as gold standard to compare existing outcome assessment tools.

Differentiating Merkel cell carcinoma of lymph nodes without a detectable primary skin tumor from other metastatic neuroendocrine carcinomas: The ELECTHIP criteria

Merkel cell carcinoma (MCC) can present as a cutaneous tumor or a lymph node metastasis without a primary tumor. MCC presenting without a primary tumor (MCCWOPT) can be misinterpreted on histologic examination as lymph node metastasis (LNM) from another neuroendocrine carcinoma (LNMNEC). However, this distinction is crucial for therapeutic management. To determine the discriminative criteria for the differential diagnosis of MCCWOPT, LNM from cutaneous MCC, and LNMNECs. Clinical, morphologic, and immunohistochemical data (expression of cytokeratins AE1, AE3, 7, 19, and 20; chromogranin A, synaptophysin, thyroid transcription factor-1 [TTF-1]), as well as the presence of Merkel cell polyomavirus (by immunohistochemistry and PCR) were compared in patients with MCCWOPT (n=17), LNM from a cutaneous MCC (n=11), and LNMNEC (n=20; 8 lung, 7 thyroid, 3 digestive tract, 2 other). MCC (including MCCWOPT and LNM from a cutaneous MCC) differed from LNMNEC by 7 discriminative criteria: 1) elderly age, 2) location of the tumor, 3) extent of the disease, 4) cytokeratin expression, 5) TTF-1 expression, 6) histologic type, and 7) Merkel cell polyomavirus detection, summarized under the acronym ELECTHIP. All MCC patients had ≥5 of the ELECTHIP criteria, whereas all patients with LNMNEC (except 1) had <3 criteria. The discriminant ability of the ELECTHIP criteria should be validated in a second independent set. MCCWOPT can be distinguished from other LNMNEC by the ELECTHIP criteria.

Index finger motor imagery EEG pattern recognition in BCI applications using dictionary cleaned sparse representation-based classification for healthy people.

Electroencephalogram (EEG)-based motor imagery (MI) brain-computer interface (BCI) has shown its effectiveness for the control of rehabilitation devices designed for large body parts of the patients with neurologic impairments. In order to validate the feasibility of using EEG to decode the MI of a single index finger and constructing a BCI-enhanced finger rehabilitation system, we collected EEG data during right hand index finger MI and rest state for five healthy subjects and proposed a pattern recognition approach for classifying these two mental states. First, Fisher's linear discriminant criteria and power spectral density analysis were used to analyze the event-related desynchronization patterns. Second, both band power and approximate entropy were extracted as features. Third, aiming to eliminate the abnormal samples in the dictionary and improve the classification performance of the conventional sparse representation-based classification (SRC) method, we proposed a novel dictionary cleaned sparse representation-based classification (DCSRC) method for final classification. The experimental results show that the proposed DCSRC method gives better classification accuracies than SRC and an average classification accuracy of 81.32% is obtained for five subjects. Thus, it is demonstrated that single right hand index finger MI can be decoded from the sensorimotor rhythms, and the feature patterns of index finger MI and rest state can be well recognized for robotic exoskeleton initiation.

O123 Optimizing the electrodiagnostic accuracy in Guillain–Barré syndrome subtypes: Sparse linear discriminant analysis versus criteria sets

Objective To optimize the electrodiagnostic accuracy for Guillain-Barre (GBS) subtypes at the first test and define the criteria for reversible conduction failure (RCF). Methods The reference electrodiagnosis was obtained in 53 demyelinating and 45 axonal GBS patients on the basis of two serial studies and results of anti-ganglioside antibodies assay. We retrospectively employed sparse linear discriminant analysis (SLDA), a supervised statistical method of classification, two existing electrodiagnostic criteria sets (Hadden et al., 1988; Rajabally et al., 2015) and one we propose that additionally evaluates duration of motor responses, sural sparing pattern and RCF in motor and sensory nerves at the second study. Results At first study the misclassification error rates, compared to reference diagnoses, were: 15.3% for sparse SLDA, 30 % for our criteria, 45% for Rajabally’s and 48% for Hadden’s criteria. SLDA identified seven most powerful electrophysiological variables differentiating demyelinating and axonal subtypes and assigned to each patient the diagnostic probabilityof belonging to either subtype. At second test 46.6% of axonal GBS patients showed RCF in two motor and 8.8% in two sensory nerves. Discussion and conclusions Basing on a single test, SLDA showed the highest diagnostic accuracy. Anyway RCF is present in a considerable percentage of axonal GBS patients and can be assessed only with serial tests. Significance Algorithms for SLDA are already available in several statistical packages and even the electromyographic machines could be easily equipped with the appropriated algorithms providing the physician with a decision support system for the diagnosis of neuropathies.

A Sparse Representation Model Using the Complete Marginal Fisher Analysis Framework and Its Applications to Visual Recognition

This paper presents an innovative sparse representation model using the complete marginal Fisher analysis (CMFA) framework for different challenging visual recognition tasks. First, a complete marginal Fisher analysis method is presented by extracting the discriminatory features in both the column space of the local samples based within the class scatter matrix and the null space of its transformed matrix. The rationale of extracting features in both spaces is to enhance the discriminatory power by further utilizing the null space, which is not accounted for in the marginal Fisher analysis method. Second, a discriminative sparse representation model is proposed by integrating a representation criterion such as the sparse representation and a discriminative criterion for improving the classification capability. In this model, the largest step size for learning the sparse representation is derived to address the convergence issues in optimization, and a dictionary screening rule is presented to purge the dictionary items with null coefficients for improving the computational efficiency. Experiments on some challenging visual recognition tasks using representative datasets, such as the Painting-91 dataset, the 15 scene categories dataset, the MIT-67 indoor scenes dataset, the Caltech 101 dataset, the Caltech 256 object categories dataset, the AR face dataset, and the extended Yale B dataset, show the feasibility of the proposed method.

Region-based Active Contour Segmentation Model with Local Discriminant Criterion

Region-based Active Contour Segmentation Model with Local Discriminant Criterion

Detecting ship targets in spaceborne infrared image based on modeling radiation anomalies

Using infrared imaging sensors to detect ship target in the ocean environment has many advantages compared to other sensor modalities, such as better thermal sensitivity and all-weather detection capability. We propose a new ship detection method by modeling radiation anomalies for spaceborne infrared image. The proposed method can be decomposed into two stages, where in the first stage, a test infrared image is densely divided into a set of image patches and the radiation anomaly of each patch is estimated by a Gaussian Mixture Model (GMM), and thereby target candidates are obtained from anomaly image patches. In the second stage, target candidates are further checked by a more discriminative criterion to obtain the final detection result. The main innovation of the proposed method is inspired by the biological mechanism that human eyes are sensitive to the unusual and anomalous patches among complex background. The experimental result on short wavelength infrared band (1.560–2.300μm) and long wavelength infrared band (10.30–12.50μm) of Landsat-8 satellite shows the proposed method achieves a desired ship detection accuracy with higher recall than other classical ship detection methods.