scholarly journals Classification of complex Wishart matrices with a diffusion–reaction system guided by stochastic distances

2015 ◽  
Vol 373 (2056) ◽  
pp. 20150118 ◽  
Author(s):  
Luis Gomez ◽  
Luis Alvarez ◽  
Luis Mazorra ◽  
Alejandro C. Frery
Keyword(s):  
Reaction System ◽  
Classification Problem ◽  
Diffusion Reaction ◽  
Speckle Reduction ◽  
Discrimination Power ◽  
Diffusion Term ◽  
Polarimetric Synthetic Aperture Radar ◽  
Positive Weights ◽  
Set Up

We propose a new method for polarimetric synthetic aperture radar (PolSAR) imagery classification based on stochastic distances in the space of random matrices obeying complex Wishart distributions. Given a collection of prototypes and a stochastic distance d (.,.), we classify any random matrix X using two criteria in an iterative set-up. First, we associate X with the class which minimizes the weighted stochastic distance w m d ( X , Z m ), where the positive weights w m are computed to max- imize the class discrimination power. Second, we improve the result by embedding the classification problem into a diffusion–reaction partial differential system where the diffusion term smooths the patches within the image, and the reaction term tends to move the pixel values towards the closest class prototype. In particular, the method inherits the benefits of speckle reduction by diffusion-like methods. Results on synthetic and real PolSAR data show the performance of the method.

scholarly journals IMPACT OF POLARIMETRIC SAR SPECKLE REDUCTION ON CLASSIFICATION OF AGRICULTURE LANDS

2019 ◽  
Vol XLII-4/W18 ◽  
pp. 379-385
Author(s):  
R. Farhadiani ◽  
S. Homayouni ◽  
A. Safari
Keyword(s):  
Speckle Reduction ◽  
Support Vector ◽  
Svm Classifier ◽  
Comparison Study ◽  
Polarimetric Synthetic Aperture Radar ◽  
Processing Step ◽  
Reduction Methods ◽  
Quantitative Indicators ◽  
The Impact

Abstract. Presence of speckle in the Polarimetric Synthetic Aperture Radar (PolSAR) images could decrease the performance of information extraction applications such as classification, segmentation, change detection, etc. Hence, an essential pre-processing step named de-speckling is needed to suppress this granular noise-like phenomenon from the PolSAR images. In this paper, a comparison study is conducted between several new PolSAR speckle reduction methods such as POSSC, PNGF, and ANLM. For this comparison, a 4-look L-band AIRSAR NASA/JPL PolSAR dataset that obtained over an agriculture land from Flevoland, Netherlands, was employed. The de-speckling assessment was completed based on some no-reference quantitative indicators. All the de-speckling methods were evaluated in terms of speckle reduction form homogeneous areas, details, and radiometric preservation, and retaining the polarimetric information. Furthermore, the impact of PolSAR de-speckling on classification was evaluated. For this purpose, Support Vector Machine (SVM) classifier was used to classify H/A/Alpha decomposition. Experimental results showed that the ANLM method was better to suppress the speckle, followed by the PNGF method. Also, the classification results showed that a proper PolSAR de-speckling could effectively increase the classification accuracy. The improvement of the Overall Accuracy based on de-speckling using the ANLM method was approximately 22% and 13% higher than the POSSC and PNGF methods, respectively.

Fingerprint of PGI Mantova Cucumis melo by ICP-MS and Chemometric Analysis

2020 ◽  
Vol 17 (1) ◽  
pp. 94-104
Author(s):  
Antonio F. Mottese ◽  
Maria R. Fede ◽  
Francesco Caridi ◽  
Giuseppe Sabatino ◽  
Giuseppe Marcianò ◽  
...  
Keyword(s):  
Cucumis Melo ◽  
Principal Component ◽  
Inorganic Elements ◽  
Discrimination Power ◽  
Icp Ms ◽  
High Discrimination ◽  
Strong Relation ◽  
Trace Element Contents ◽  
Safe Food

Background and Objectives: In this work, yellow and green varieties of Cucumis melo fruits belonging to different cultivars were studied. In detail, three Sicilian cultivars of winter melons tutelated by TAP (Traditional agro-alimentary products) labels were considered, whereas asun protected the Calabrian winter melon was studied too. With the aim to compare the selective uptakes of inorganic elements among winter and summer fruits, the “PGI Melone Mantovano” was investigated. The purpose of this work was to apply the obtained results i) to guarantee the quality and healthiness of fruits, ii) to producers defend, iii) to help the customers in safe food purchase. Method: All samples were analyzed by ICP-MS and the obtained results, subsequently, were subjected to Cluster analysis (CA), Principal component analysis (PCA) and Canonical discriminant analysis (CDA). Results: CA results were generally in agreement with samples origin, whereas the PCA elaboration has confirmed the presence of a strong relation between fruit origins and trace element contents. In particular, two principal components justified the 57.32% of the total variance (PC1= 40.95%, PC2= 16.37%). Finally, the CDA approach has provided several functions with high discrimination power, confirmed by the correct classification of all samples (100%). Conclusions: CA, PCA and CDA could represent an integrated to label to discriminate the origin of agri-food products and, thus, protect and guarantee their healthiness.

scholarly journals An aggregate method for thorax diseases classification

2021 ◽  
Vol 11 (1) ◽  
Author(s):  
Bayu Adhi Nugroho
Keyword(s):  
Network Architecture ◽  
Classification Problem ◽  
Calculation Algorithm ◽  
Training Pattern ◽  
Deep Network ◽  
Network Training ◽  
Chest X Ray ◽  
Medical Image Classification ◽  
Positive Pattern

AbstractA common problem found in real-word medical image classification is the inherent imbalance of the positive and negative patterns in the dataset where positive patterns are usually rare. Moreover, in the classification of multiple classes with neural network, a training pattern is treated as a positive pattern in one output node and negative in all the remaining output nodes. In this paper, the weights of a training pattern in the loss function are designed based not only on the number of the training patterns in the class but also on the different nodes where one of them treats this training pattern as positive and the others treat it as negative. We propose a combined approach of weights calculation algorithm for deep network training and the training optimization from the state-of-the-art deep network architecture for thorax diseases classification problem. Experimental results on the Chest X-Ray image dataset demonstrate that this new weighting scheme improves classification performances, also the training optimization from the EfficientNet improves the performance furthermore. We compare the aggregate method with several performances from the previous study of thorax diseases classifications to provide the fair comparisons against the proposed method.

scholarly journals Implementation of a deep learning model for automated classification of Aedes aegypti (Linnaeus) and Aedes albopictus (Skuse) in real time

2021 ◽  
Vol 11 (1) ◽  
Author(s):  
Song-Quan Ong ◽  
Hamdan Ahmad ◽  
Gomesh Nair ◽  
Pradeep Isawasan ◽  
Abdul Hafiz Ab Majid
Keyword(s):  
Deep Learning ◽  
Aedes Aegypti ◽  
Real Time ◽  
Aedes Albopictus ◽  
Automated Classification ◽  
Expert Performance ◽  
Deep Convolutional Neural Networks ◽  
Significant Difference ◽  
Set Up

AbstractClassification of Aedes aegypti (Linnaeus) and Aedes albopictus (Skuse) by humans remains challenging. We proposed a highly accessible method to develop a deep learning (DL) model and implement the model for mosquito image classification by using hardware that could regulate the development process. In particular, we constructed a dataset with 4120 images of Aedes mosquitoes that were older than 12 days old and had common morphological features that disappeared, and we illustrated how to set up supervised deep convolutional neural networks (DCNNs) with hyperparameter adjustment. The model application was first conducted by deploying the model externally in real time on three different generations of mosquitoes, and the accuracy was compared with human expert performance. Our results showed that both the learning rate and epochs significantly affected the accuracy, and the best-performing hyperparameters achieved an accuracy of more than 98% at classifying mosquitoes, which showed no significant difference from human-level performance. We demonstrated the feasibility of the method to construct a model with the DCNN when deployed externally on mosquitoes in real time.

scholarly journals Classification of unlabeled online media

2021 ◽  
Vol 11 (1) ◽  
Author(s):  
Sakthi Kumar Arul Prakash ◽  
Conrad Tucker
Keyword(s):  
Social Media ◽  
Real World ◽  
Graphical Model ◽  
Ground Truth ◽  
Classification Problem ◽  
Machine Learning Algorithms ◽  
Social Media Networks ◽  
Online Social Media ◽  
Wide Range

AbstractThis work investigates the ability to classify misinformation in online social media networks in a manner that avoids the need for ground truth labels. Rather than approach the classification problem as a task for humans or machine learning algorithms, this work leverages user–user and user–media (i.e.,media likes) interactions to infer the type of information (fake vs. authentic) being spread, without needing to know the actual details of the information itself. To study the inception and evolution of user–user and user–media interactions over time, we create an experimental platform that mimics the functionality of real-world social media networks. We develop a graphical model that considers the evolution of this network topology to model the uncertainty (entropy) propagation when fake and authentic media disseminates across the network. The creation of a real-world social media network enables a wide range of hypotheses to be tested pertaining to users, their interactions with other users, and with media content. The discovery that the entropy of user–user and user–media interactions approximate fake and authentic media likes, enables us to classify fake media in an unsupervised learning manner.

scholarly journals Rapid Classification of Clostridioides difficile Strains Using MALDI-TOF MS Peak-Based Assay in Comparison with PCR-Ribotyping

2021 ◽  
Vol 9 (3) ◽  
pp. 661
Author(s):  
Adriana Calderaro ◽  
Mirko Buttrini ◽  
Monica Martinelli ◽  
Benedetta Farina ◽  
Tiziano Moro ◽  
...  
Keyword(s):  
Ionization Time ◽  
Maldi Tof Ms ◽  
Maldi Tof ◽  
Flight Mass Spectrometry ◽  
Clostridioides Difficile ◽  
Pcr Ribotyping ◽  
Typing Methods ◽  
Set Up ◽  
Tof Ms

Typing methods are needed for epidemiological tracking of new emerging and hypervirulent strains because of the growing incidence, severity and mortality of Clostridioides difficile infections (CDI). The aim of this study was the evaluation of a typing Matrix-Assisted Desorption/Ionization-Time of Flight Mass Spectrometry (MALDI-TOF MS (T-MALDI)) method for the rapid classification of the circulating C. difficile strains in comparison with polymerase chain reaction (PCR)-ribotyping results. Among 95 C. difficile strains, 10 ribotypes (PR1–PR10) were identified by PCR-ribotyping. In particular, 93.7% of the isolates (89/95) were grouped in five ribotypes (PR1–PR5). For T-MALDI, two classifying algorithm models (CAM) were tested: the first CAM involved all 10 ribotypes whereas the second one only the PR1–PR5 ribotypes. Better performance was obtained using the second CAM: recognition capability of 100%, cross-validation of 96.6% and agreement of 98.4% (60 correctly typed strains, limited to PR1–PR5 classification, out of 61 examined strains) with PCR-ribotyping results. T-MALDI seems to represent an alternative to PCR-ribotyping in terms of reproducibility, set up time and costs, as well as a useful tool in epidemiological investigation for the detection of C. difficile clusters (either among CAM included ribotypes or out-of-CAM ribotypes) involved in outbreaks.

VirusTaxo: Taxonomic classification of virus genome using multi-class hierarchical classification by k-mer enrichment

2021 ◽  
Author(s):  
Rajan Saha Raju ◽  
Abdullah Al Nahid ◽  
Preonath Shuvo ◽  
Rashedul Islam
Keyword(s):  
Genome Sequence ◽  
Rna Viruses ◽  
Hierarchical Classification ◽  
Classification Problem ◽  
Virus Genome ◽  
Taxonomic Classification ◽  
Dna Viruses ◽  
Dna And Rna ◽  
Full Length Genome

AbstractTaxonomic classification of viruses is a multi-class hierarchical classification problem, as taxonomic ranks (e.g., order, family and genus) of viruses are hierarchically structured and have multiple classes in each rank. Classification of biological sequences which are hierarchically structured with multiple classes is challenging. Here we developed a machine learning architecture, VirusTaxo, using a multi-class hierarchical classification by k-mer enrichment. VirusTaxo classifies DNA and RNA viruses to their taxonomic ranks using genome sequence. To assign taxonomic ranks, VirusTaxo extracts k-mers from genome sequence and creates bag-of-k-mers for each class in a rank. VirusTaxo uses a top-down hierarchical classification approach and accurately assigns the order, family and genus of a virus from the genome sequence. The average accuracies of VirusTaxo for DNA viruses are 99% (order), 98% (family) and 95% (genus) and for RNA viruses 97% (order), 96% (family) and 82% (genus). VirusTaxo can be used to detect taxonomy of novel viruses using full length genome or contig sequences.AvailabilityOnline version of VirusTaxo is available at https://omics-lab.com/virustaxo/.

Using transfer learning for diabetic retinopathy stage classification

2021 ◽  
Vol ahead-of-print (ahead-of-print) ◽  
Author(s):  
Enas M.F. El Houby
Keyword(s):  
Diabetic Retinopathy ◽  
Transfer Learning ◽  
Grade Level ◽  
Design Methodology ◽  
Content Type ◽  
Stage Classification ◽  
Set Up ◽  
Retinal Fundus ◽  
Automated Methods

PurposeDiabetic retinopathy (DR) is one of the dangerous complications of diabetes. Its grade level must be tracked to manage its progress and to start the appropriate decision for treatment in time. Effective automated methods for the detection of DR and the classification of its severity stage are necessary to reduce the burden on ophthalmologists and diagnostic contradictions among manual readers.Design/methodology/approachIn this research, convolutional neural network (CNN) was used based on colored retinal fundus images for the detection of DR and classification of its stages. CNN can recognize sophisticated features on the retina and provides an automatic diagnosis. The pre-trained VGG-16 CNN model was applied using a transfer learning (TL) approach to utilize the already learned parameters in the detection.FindingsBy conducting different experiments set up with different severity groupings, the achieved results are promising. The best-achieved accuracies for 2-class, 3-class, 4-class and 5-class classifications are 86.5, 80.5, 63.5 and 73.7, respectively.Originality/valueIn this research, VGG-16 was used to detect and classify DR stages using the TL approach. Different combinations of classes were used in the classification of DR severity stages to illustrate the ability of the model to differentiate between the classes and verify the effect of these changes on the performance of the model.

scholarly journals The Algebraic and Geometric Classification of Four Dimensional Superalgebras

2021 ◽  
Author(s):  
◽  
Aaron Armour
Keyword(s):  
Classification Problem ◽  
Associative Algebras ◽  
New Variety ◽  
Graded Algebras ◽  
Finite Representation ◽  
Algebraic Classification ◽  
Irreducible Components ◽  
Module Algebras ◽  
Sweedler’S Hopf Algebra

<p><b>The algebraic and geometric classification of k-algbras, of dimension fouror less, was started by Gabriel in “Finite representation type is open” [12].</b></p> <p>Several years later Mazzola continued in this direction with his paper “Thealgebraic and geometric classification of associative algebras of dimensionfive” [21]. The problem we attempt in this thesis, is to extend the resultsof Gabriel to the setting of super (or Z2-graded) algebras — our main effortsbeing devoted to the case of superalgebras of dimension four. Wegive an algebraic classification for superalgebras of dimension four withnon-trivial Z2-grading. By combining these results with Gabriel’s we obtaina complete algebraic classification of four dimensional superalgebras.</p> <p>This completes the classification of four dimensional Yetter-Drinfeld modulealgebras over Sweedler’s Hopf algebra H4 given by Chen and Zhangin “Four dimensional Yetter-Drinfeld module algebras over H4” [9]. Thegeometric classification problem leads us to define a new variety, Salgn —the variety of n-dimensional superalgebras—and study some of its properties.</p> <p>The geometry of Salgn is influenced by the geometry of the varietyAlgn yet it is also more complicated, an important difference being thatSalgn is disconnected. While we make significant progress on the geometricclassification of four dimensional superalgebras, it is not complete. Wediscover twenty irreducible components of Salg4 — however there couldbe up to two further irreducible components.</p>

Sign in / Sign up

Export Citation Format

Share Document