Search

Remote sensing image retrieval (RSIR) may be a fundamental task in remote sensing. Most content-based image retrieval (CBRSIR) approaches take an easy distance as similarity criteria. A retrieval method supported weighted distance and basic features of Convolutional Neural Network (CNN) is proposed during this letter. the strategy contains two stages. First, in offline stage, the pretrained CNN will be fine-tuned by some labelled images from our target data set, then accustomed extract CNN features, and labelled the pictures within the retrieval data set. Second, in online stage, we extract features of the query image by using fine-tuned CNN model and calculate the load of every image class and apply them to calculate the space between the query image and also the retrieved images. Experiments and methods are conducted on two Remote Sensing Image Retrieval data sets. Compared with the state-of the-art methods, the proposed method significantly improves retrieval performance.

In this paper, we propose a novel approach for outlier detection, called local projections, which is based on concepts of Local Outlier Factor (LOF) (Breunig et al., 2000) and RobPCA (Hubert et al., 2005). By using aspects of both methods, our algorithm is robust towards noise variables and is capable of performing outlier detection in multi-group situations. We are further not reliant on a specific underlying data distribution. For each observation of a dataset, we identify a local group of dense nearby observations, which we call a core, based on a modification of the k-nearest neighbours algorithm. By projecting the dataset onto the space spanned by those observations, two aspects are revealed. First, we can analyze the distance from an observation to the center of the core within the projection space in order to provide a measure of quality of description of the observation by the projection. Second, we consider the distance of the observation to the projection space in order to assess the suitability of the core for describing the outlyingness of the observation. These novel interpretations lead to a univariate measure of outlyingness based on aggregations over all local projections, which outperforms LOF and RobPCA as well as other popular methods like PCOut (Filzmoser et al., 2008) and subspace-based outlier detection (Kriegel et al., 2009) in our simulation setups. Experiments in the context of real-word applications employing datasets of various dimensionality demonstrate the advantages of local projections.

Image segmentation is the basis of image analysis, object tracking, and other fields. However, image segmentation is still a bottleneck due to the complexity of images. In recent years, fuzzy clustering is one of the most important selections for image segmentation, which can retain information as much as possible. However, fuzzy clustering algorithms are sensitive to image artifacts. In this study, an improved image segmentation algorithm based on patch-weighted distance and fuzzy clustering is proposed, which can be divided into two steps. First, the pixel correlation between adjacent pixels is retrieved based on patch-weighted distance, and then the pixel correlation is used to replace the influence of neighboring information in fuzzy algorithms, thereby enhancing the robustness. Experiments on simulated, natural and medical images illustrate that the proposed schema outperforms other fuzzy clustering algorithms.

Convolutional neural networks (CNN) is the most mainstream solution in the field of image retrieval. Deep metric learning is introduced into the field of image retrieval, focusing on the construction of pair-based loss function. However, most pair-based loss functions of metric learning merely take common vector similarity (such as Euclidean distance) of the final image descriptors into consideration, while neglecting other distribution characters of these descriptors. In this work, we propose relative distribution entropy (RDE) to describe the internal distribution attributes of image descriptors. We combine relative distribution entropy with the Euclidean distance to obtain the relative distribution entropy weighted distance (RDE-distance). Moreover, the RDE-distance is fused with the contrastive loss and triplet loss to build the relative distributed entropy loss functions. The experimental results demonstrate that our method attains the state-of-the-art performance on most image retrieval benchmarks.

Supervoxel segmentation has become an essential tool to medical image analysis for three-dimension MR image. However, in no consideration of the intensity inhomogeneity in 2D/3D MR image, the state-of-the-art supervoxel segmentation methods do not satisfy the further analysis, such as tissue classification according to intensity feature. In order to overcome the above-mentioned issues, we propose a modified supervoxel segmentation method for three-dimension MR image, which integrates the bias field into the weighted distance metric to determine the nearest cluster center. The supervoxel segmentation and bias correction can be simultaneously completed in our method. Especially, the bias corrected image lays the foundation for the supervoxel classification in accordance with the intensity feature. The experimental results and quantitative evaluation showed that the supervoxels obtained by our method are adherence to the MR tissue boundaries, and the bias corrected image is positive for the intensity feature extraction.

Distance and similarity measures are very important topics in fuzzy set theory. In this paper, we propose some new distance measures for generalized hesitant fuzzy sets. We investigate the connections of the aforementioned distance measures and further develop a number of generalized hesitant ordered weighted distance measures.

An original method for multi-fault detection in synchronous machine is proposed in this paper. This method aims to answer two questions: how to detect a fault when only the normal functioning is known? How to differentiate two different faults from one fault with two severities? The proposed method relies on the use of a recursive Principal Components Analysis (PCA), which is updated each time a new fault is detected. The detection is based on a weighted distance criteria that takes into account the contribution of the different components. A geometrical criteria is also proposed to differentiate new faults from existing ones. The method has been successfully tested on a simulation database of motor currents with different levels of unbalance and eccentricity. It is then tested on real data from a 5.5 kW synchronous machine with three different levels of unbalance.

In multi-label classification, each instance is associated with a set of pre-specified labels. One common approach is to use Binary Relevance (BR) paradigm to learn each label by a base classifier separately. Use of k-Nearest Neighbor (kNN) as the base classifier (denoted as BRkNN) is a simple, descriptive and powerful approach. In binary relevance a highly imbalanced view of dataset is used. However, kNN is known to perform poorly on imbalanced data. One approach to deal with this is to define the distance function in a parametric form and use the training data to adjust the parameters (i.e. adjusting boundaries between classes) by optimizing a performance measure customized for imbalanced data e.g. F -measure. Prototype Weighting (PW) scheme presented in the literature (Paredes & Vidal, 2006) uses gradient descent to specify the parameters by minimizing the classification error-rate on training data. This paper presents a generalized version of PW. First, instead of minimizing the error-rate proposed in PW, the generalized PW supports also other objective functions that use elements of confusion matrix (including F -measure). Second, PW originally presented for 1NN is extended to the general case of kNN (i.e., k > = 1 ). For problems having highly overlapped classes, it is expected to perform better since a value of k > 1 produces smoother decision boundaries which in turn can improve generalization. In multi-label problems with many labels or problems with highly overlapped classes, the proposed generalized PW is expected to significantly improve the performance as it involves many decision boundaries. The performance of the proposed method has been compared with state-of-the-art methods in multi-label classification containing 6 lazy classifiers based on kNN. Experiments show that the proposed method significantly outperforms other methods.

Plant leaf identification is a significant challenge in the fields of computer vision and pattern recognition. This article presents a new approach to plant leaf identification, one that integrates shape and texture characteristics. First, we introduce the shape and texture features used by the proposed plant leaf recognition method. The proposed multiscale triangle descriptor (MTD) is employed to characterize the shape information of a plant leaf, and the local binary pattern histogram Fourier (LBP-HF) is used as the texture feature. Then, the shape and texture features of a leaf image are combined by weighted distance measurement, where L 1 distance and chi-square distance are used for shape and texture features, respectively. The proposed approach provides a robust descriptor for the task of plant leaf recognition by combining the complementary MTD and LBP-HF features. The proposed approach has been thoroughly evaluated on three benchmark leaf datasets, including the Flavia, Swedish and MEW2012 leaf datasets. Our method achieves 77.6%, 85.7%, and 67.5% retrieval accuracy on the Flavia, Swedish and MEW2012 leaf datasets, respectively, while the corresponding classification accuracy is 99.1%, 98.4%, 95.6%. The recognition performance of our method is better or comparable to prior state-of-the-art plant leaf recognition method.

Existing distance-based outlier mining methods do not consider the impact of each attribute's importance degree, thereby resulting in poor mining accuracies. To address this problem, we propose a new outlier mining algorithm – Miner* – that makes use of information entropy and Weighted Distance Sum to substantially improve mining accuracies. Miner* employs information entropy to determine weight values indicating the importance degrees of data attributes. An input dataset is reduced by Miner* through the neighbour-radius-based pruning technologies. Thus, Miner* obtains a candidate outlier set by removing any data objects that are unlikely to be outliers. Miner* calculates the weighted distance sum value Wkof each object in the candidate outlier set; Wkvalue ranks the top n to be regarded as outliers. Due to the sum of distance, which takes full advantage of the clustering characteristics of the dataset, edge distribution data objects and local outliers can be effectively mined out. To demonstrate the effectiveness of the Miner* algorithm, we implement Miner* in a prototype system to detect star spectrum data objects with abnormal characteristic lines. Our experimental results show that the algorithm in Miner* achieves high accuracy, high scalability, and low man-made influence by utilizing UCI and star spectrum dataset. Our results also confirm that Miner* is feasible and effective in mining spectrum data with abnormal characteristic lines from massive star spectrum dataset.

Distance and similarity measures are very important topics in fuzzy set theory. In this paper, we propose a variety of distance measures for generalized hesitant fuzzy sets, based on which the corresponding similarity measures can be obtained. We investigate the connections of the mentioned distance measures and further develop a number of generalized hesitant ordered weighted distance measures and generalized hesitant ordered weighted similarity measures.

As a molecular and noninvasive detection technology, Raman spectroscopy is promising for use in the early diagnosis of tumors. The SNR of spectra obtained from portable Raman spectrometers is low, which makes classification more difficult. A classification algorithm with a high recognition rate is required. In this paper, an algorithm of entropy weighted local-hyperplane k-nearest-neighbor (EWHK) is proposed for the identification of the spectra. When calculating the weighted distance between the prediction and the sample hyperplane, EWHK introduces the information entropy weighting to improve the algorithm of adaptive weighted k-local hyperplane (AWKH). It can reflect all of the sample information in the classification objectively and improve the classification accuracy. The breast cancer detection experimental results of EWHK showed a significant improvement compared with those of AWKH and k-nearest neighbor (KNN). The EWHK classifier yielded an average diagnostic accuracy of 92.33%, a sensitivity of 93.81%, a specificity of 87.77%, a positive prediction rate of 95.99% and a negative prediction rate of 83.69% during randomized grouping validation. The algorithm is effective for cancer diagnosis.

Follicle quantification refers to the computation of the number and size of follicles in 3D ultrasound volumes of the ovary. This is one of the key factors in determining hormonal dosage during female infertility treatments. In this paper, we propose an automated algorithm to detect and segment follicles in 3D ultrasound volumes of the ovary for quantification. In a first of its kind attempt, we employ noise-robust phase symmetry feature maps as likelihood function to perform mean-shift based follicle center detection. Max-flow algorithm is used for segmentation and gray weighted distance transform is employed for post-processing the results. We have obtained state-of-the-art results with a true positive detection rate of >90% on 26 3D volumes with 323 follicles.

The signature forgery can be restricted by either online or offline signature verification techniques. It verifies the signature by performing a match with the pre-processed signature dynamically by detecting the motion of stylus during signature while on other hand, offline verifies by performing a match using the two dimensional scanned image of the signature. This paper studies about the various techniques available in offline signature verification along with their shadows.

Person re-identification remains a challenging problem due to large variations of poses, occlusions, illumination and camera views. To learn both feature representation and similarity metric simultaneously, deep metric learning methods using triplet convolutional neural network have been applied in person re-identification. In this paper, we propose a body structure based triplet convolutional neural network (BSTCNN) for person re-identification. Specifically, a four-branch CNN architecture is built to learn features from different body parts. Body-part features are then fused in score level with a novel weighted distance layer which learns weights for each body part. We further design an improved triplet loss function to speed up convergence and boost the performance. Experimental results on two challenging datasets (CUHK01 and PRID2011) demonstrate that our approach significantly outperforms the state-of-the-art methods.

As a nonparametric classification algorithm, K-Nearest Neighbor (KNN) is very efficient and can be easily realized. However, for large dataset, the computational demands for classifying instances using KNN can be expensive. A way to solve this problem is through the condensing approach. It means we remove instances that will bring computational burden but do not contribute to better classification accuracy. This paper proposes a novel weighted distance KNN algorithm based on instances condensing algorithm. The proposed idea is to extract some representative instances and take the processed result as a new training sample set. Meanwhile, use the distance-weighted WDKNN algorithm to improve the prediction accuracy, our experiments show that the proposed strategy can dramatically shorten the time consumption compared with the traditional KNN. On average, the speedup ratios improve 90% while classification accuracy only has 2% decreases.

Clinical data analysis is of fundamental importance, as classifications and detailed characterizations of diseases help physicians decide suitable management for patients, individually. In our study, we adopt diffusion maps to embed the data into corresponding lower dimensional representation, which integrate the information of potentially nonlinear progressions of the diseases. To deal with nonuniformaity of the data, we also consider an alternative distance measure based on the estimated local density. Performance of this modification is assessed using artificially generated data. Another clinical dataset that comprises metabolite concentrations measured with magnetic resonance spectroscopy was also classified. The algorithm shows improved results compared with conventional Euclidean distance measure.

Electroencephalogram (EEG) signals are often used to diagnose diseases such as seizure, alzheimer, and schizophrenia. One main problem with the recorded EEG samples is that they are not equally reliable due to the artifacts at the time of recording. EEG signal classification algorithms should have a mechanism to handle this issue. It seems that using adaptive classifiers can be useful for the biological signals such as EEG. In this paper, a general adaptive method named weighted distance nearest neighbor (WDNN) is applied for EEG signal classification to tackle this problem. This classification algorithm assigns a weight to each training sample to control its influence in classifying test samples. The weights of training samples are used to find the nearest neighbor of an input query pattern. To assess the performance of this scheme, EEG signals of thirteen schizophrenic patients and eighteen normal subjects are analyzed for the classification of these two groups. Several features including, fractal dimension, band power and autoregressive (AR) model are extracted from EEG signals. The classification results are evaluated using Leave one (subject) out cross validation for reliable estimation. The results indicate that combination of WDNN and selected features can significantly outperform the basic nearest-neighbor and the other methods proposed in the past for the classification of these two groups. Therefore, this method can be a complementary tool for specialists to distinguish schizophrenia disorder.

 Abstract—The datasets used in many real applications are highly imbalanced which makes classification problem hard. Classifying the minor class instances is difficult due to bias of the classifier output to the major classes. Nearest neighbor is one of the most popular and simplest classifiers with good performance on many datasets. However, correctly classifying the minor class is commonly sacrificed to achieve a better performance on others. This paper is aimed to improve the performance of nearest neighbor in imbalanced domains, without disrupting the real data distribution. Prototype-weighting is proposed, here, to locally adapting the distances to increase the chance of prototypes from minor class to be the nearest neighbor of a query instance. The objective function is, here, G-mean and optimization process is performed using gradient ascent method. Comparing the experimental results, our proposed method significantly outperformed similar works on 24 standard data sets.

Plastic surgery provide a way to enhance the facial appearance. The non-linear variations introduced by the plastic surgery has raised a challenge for face recognition algorithms. In this research we match the face image before and after the plastic surgery. First generate non-disjoint face granules at multiple levels of granularity. The feature extractors are used to extract features from the face granules. The features are then processed by using principal component analysis (PCA) algorithm. Evaluate the weighted distance and match the pre and post surgery images based on weighted distance. The proposed system yield high identification accuracy and take less time for recognition as compared to the existing system.

Traditional PSO(Particle Swarm Optimization) algorithm has the problems of particle cross-border and premature convergence while solving the normalized constrained optimization problem. Our paper uses the attractor and spatial zoom method, and presented the normalized PSO algorithm. Secondly, each attribute is treated equally in the traditional classification algorithm, without considering the differences in attribute measure and contribution, which cause the problem of low classification accuracy. Our paper introduce the use of the normalized PSO algorithm to solve the optimal attribute normalized weighted distance. For example, in KNN classifier, leave-one-out experimental results with multiple UCI data sets show that the classification accuracy using normalization PSO algorithm to calculate normalized weighted distance is higher than using PSO algorithm and traditional none-attribute weighted classifiers.

Automatic image annotation is a challenging problem due to the label-image-matching, label-imbalance and label-missing problems. Some research tried to address part of these problems but didn't integrate them. In this paper, an adaptive weighted distance method which incorporates the CNN convolutional neural network feature and multiple handcrafted features is proposed to handle the label-image-matching and label-imbalance issues, while the K nearest neighbors framework is improved by using the neighborhood with all labels which can reduce the effects of the label-missing problem. Finally, experiments on three benchmark datasets Corel-5k, ESP-Game and IAPRTC-12 for image annotation are performed, and the results show that our approach is competitive to the state-of-the-art methods.

A natural methodology for discriminating functional data is based on the distances from the observation or its derivatives to group representative functions (usually the mean) or their derivatives. It is proposed to use a combination of these distances for supervised classification. Simulation studies show that this procedure performs very well, resulting in smaller testing classification errors. Applications to real data show that this technique behaves as well as-and in some cases better than-existing supervised classification methods for functions.

In this paper, we propose a variety of distance measures for hesitant fuzzy sets, based on which the corresponding similarity measures can be obtained. We investigate the connections of the aforementioned distance measures and further develop a number of hesitant ordered weighted distance measures and hesitant ordered weighted similarity measures. They can alleviate the influence of unduly large (or small) deviations on the aggregation results by assigning them low (or high) weights. Several numerical examples are provided to illustrate these distance and similarity measures.

By terms of extracting quantization values of each index making contributions to classification, this paper defines index classification weight; and also defines class representative points, weighted distance between samples and representative points; provides an iterative algorithm of searching class representative points, establishes a supervised clustering method based on representative points and it is apply into Fault diagnosis of Diesel Engine.

Aiming at the feature of signer-independent sign language recognition the training data complexity caused by mass data and noticeable distinctions between different people data, the weighted KNN/HMM model is presented in this paper. This model is made of two blocks, which is part of sign language classification and recognition. In classing part, the KNN (K-Nearest Neighbor, KNN) is used to learn the training samples. Considering the different contributions of sign language features to pattern classification we give different weight to different characteristics. And the category of test sample is decided by the sum of weighted distance. In recognition part, weighted KNN classification result is taken as the state-input of HMM (Hidden Markov models, HMM) to implement sign language recognition, combine with the ability to temporal data modeling and fuzzy inference of HMM model. Experiment results show that weighted KNN/HMM sign language recognition model are efficient on either recognition speed or recognition rate.

This paper aims at a dynamic gesture recognition method based on the improved DTW algorithm. Firstly, the 3D position of human skeletal points is obtained through the analysis of the depth information, which is obtained by SDK of the Kinect sensor. 8 points are selected as the hand movement characteristics, and the mathematical model of the gesture is established by the method of weighted distance. Then the DTW algorithm is improved by distortion threshold and the path constraints, which is used for training templates and dynamic gesture recognition. The results show that this method can realize the dynamic gesture recognition and have a good real-time performance and robustness. At the same time, the improved DTW algorithm has great improvement in speed and accuracy.

International audience; The performance of nearest-neighbor (NN) classifiers is known to be very sensitive to the distance metric used in classifying a query pattern, especially in scarce-prototype cases. In this paper, a class-conditional weighted (CCW) distance metric related to both the class labels of the prototypes and the query patterns is proposed. Compared with the existing distance metrics, the proposed metric provides more flexibility to design the feature weights so that the local specifics in feature space can be well characterized. Based on the proposed CCW distance metric, a multi-hypothesis nearest-neighbor (MHNN) classifier is developed. The scheme of the proposed MHNN classifier is to classify the query pattern under multiple hypotheses in which the nearest-neighbor sub-classifiers can be implemented based on the CCW distance metric. Then the classification results of multiple sub-classifiers are combined to get the final result. Under this general scheme, a specific realization of the MHNN classifier is developed within the framework of Dempster-Shafer theory due to its good capability of representing and combining uncertain information. Two experiments based on synthetic and real data sets were carried out to show the effectiveness of the proposed technique.

Traditional point-by-point image similarity metrics, such as the l 2 -norm, are not always consistent with human perception, especially in textured regions. We consider the problem of identifying textures that are perceptually identical to a query texture; this is important for image retrieval, compression, and restoration applications. Recently proposed structural texture similarity (STSIM) metrics assign high similarity scores to such perceptually identical textured patches, even though they may have significant pixel-wise deviations. We use an STSIM approach that compares a set of statistical patch descriptors through a weighted distance, and, given a dataset of labeled texture images partitioned into classes of perceptually identical patches, we calculate the weights as the variances of each statistic centered around the mean of its class. Experimental results demonstrate that the proposed approach outperforms existing structural similarity metrics and STSIMs as well as traditional point-by-point metrics when assessing texture similarity in both noisy and noise-free conditions.

The traditional centroid-based classifiers cannot be directly applied to categorical data classification due to the undefined concept of centroid for a categorical class, and the lack of an effective distance measure for categorical objects. In this paper, two centroid-based classifiers are proposed for categorical data classification. We propose a new formulation for the centroid of categorical classes to address the first problem, while two weighted distance measures are defined for the second problem. The experimental results conducted on real-world data sets show the effectiveness of the proposed methods.

In this paper, a new learned basis set algorithm (3D-LBT) based on 3D-DCT (Discrete Cosine Transform) is proposed for video fingerprinting and matching, in which for different video categories an Adaboost-based machine learning method is applied to each category of videos for selecting suitable sets of 3D-DCT coefficients to generate fingerprints, and a weighted distance of fingerprints is also defined for fingerprint matching. Our experimental results have illustrated that the proposed algorithm outperforms the conventional 3D-DCT algorithm and the 3D-RBT (Randomized Basis seT) algorithm in terms of robustness and uniqueness. Moreover, the proposed algorithm has better security performance for copyright applications.

The traditional K-nearest neighbor algorithm usually adopts Euclidean distance formula to measure the distance between two samples. Since each attribute functions differently in the actual sample data collection, the accuracy of the classification will be reduced consequently. In order to improve traditional KNN and KNN with weighted distance which is on the distance definition and test mode, this article proposes one method to measure the attribute value and entropy weight, namely K-nearest neighbor algorithm based on weighted entropy of attribute value. The experiment indicated that, compared with the traditional K-nearest neighbor algorithm, the algorithm proposed in this article can not only guarantee the efficiency of classification but also enhance the accuracy of classification.

Distance measures are an important means to find the difference of data. In this paper, we develop a type of hybrid weighted distance measures which are based on the weighted distance measures and the ordered weighted averaging operator, and aslo point out some of their special cases. Then, we apply the developed measures to pattern recognition.

Translation of human intentions into control signals for a computer, so called Brain-Computer Interface (BCI), has been a growing research field during the last years. In this way, classification of mental tasks is under investigation in the BCI society as a basic research. In this paper, a Weighted Distance Nearest Neighbor (WDNN) classifier is presented to improve the classification rate between the left and right imagery tasks in which a weight is assigned to each stored instance. The specified weight of each instance is then used for calculating the distance of a test pattern to that instance. We propose an iterative learning algorithm to specify the weights of training instances such that the error rate of the classifier on training data is minimized. ElectroEncephaloGram (EEG) signals are caught from four familiar subjects with the cue-based BCI. The proposed WDNN classifier is applied to the band power and fractal dimension features, which are extracted from EEG signals to classify mental tasks. Results show that our proposed method performs better in some subjects in comparison with the LDA and SVM, as well-known classifiers in the BCI field.

The performance of Nearest Neighbor (NN) classifier is highly dependent on the distance (or similarity) function used to find the NN of an input test pattern. Many of the proposed algorithms try to optimize the accuracy of the NN rule using a weighted distance function. In this scheme, a weight parameter is learned for each of the training instances. The weights of training instances are used in the generalization phase to find the NN of an input test pattern. The Weighted Distance Nearest Neighbor (WDNN) algorithm attempts to maximize the leave-one-out classification rate of the training set by adjusting the weight parameters. The procedure simply leads to weights that overfit the train data, which degrades the performance of the method especially in noisy environments. In this paper, we propose an enhanced version of WDNN, called Overfit Avoidance for WDNN (OAWDNN), that significantly outperforms the algorithm in generalization phase. The proposed method uses an early stopping approach to decrease instance weights specified by WDNN, which implicitly makes the class boundary smooth and consequently more generalized. In order to evaluate robustness of the algorithm, class label noise is added to a variety of UCI datasets. The experimental results show the supremacy of the proposed method in generalization accuracy.

In real-world applications, it has been often observed that class imbalance (significant differences in class prior probabilities) may produce an important deterioration of the classifier performance, in particular with patterns belonging to the less represented classes. This effect becomes especially significant on instance-based learning due to the use of some dissimilarity measure. We analyze the effects of class imbalance on the classifier performance and how the overlap has influence on such an effect, as well as on several techniques proposed in the literature to tackle the class imbalance. Besides, we study how these methods affect to the performance on both classes, not only on the minority class as usual.

In this paper we propose a new fuzzy linguistic method to recognize odorant samples. The method is applied to raw experimental data collected from a sensor array that comprises sixteen conducting polymer sensors with partially overlapping sensitivities. The sensors are exposed to odor samples and the percentage change in resistance is used for classification. The method describes the shape of each sensor response in terms of linguistic expressions derived from a fuzzy partition of the area occupied by the response. A purposely-defined weighted distance is used to compare the linguistic descriptions. Results on the application of the method to the classification of three chemicals in two different concentrations are presented.

BackgroundThe composition vector (CV) method has been proved to be a reliable and fast alignment-free method to analyze large COI barcoding data. In this study, we modify this method for analyzing multi-gene datasets for plant DNA barcoding. The modified method includes an adjustable-weighted algorithm for the vector distance according to the ratio in sequence length of the candidate genes for each pair of taxa.Methodology/principal findingsThree datasets, matK+rbcL dataset with 2,083 sequences, matK+rbcL dataset with 397 sequences and matK+rbcL+trnH-psbA dataset with 397 sequences, were tested. We showed that the success rates of grouping sequences at the genus/species level based on this modified CV approach are always higher than those based on the traditional K2P/NJ method. For the matK+rbcL datasets, the modified CV approach outperformed the K2P-NJ approach by 7.9% in both the 2,083-sequence and 397-sequence datasets, and for the matK+rbcL+trnH-psbA dataset, the CV approach outperformed the traditional approach by 16.7%.ConclusionsWe conclude that the modified CV approach is an efficient method for analyzing large multi-gene datasets for plant DNA barcoding. Source code, implemented in C++ and supported on MS Windows, is freely available for download at http://math.xtu.edu.cn/myphp/math/research/source/Barcode_source_codes.zip.

International audience; Human action recognition based on RGB-D sequences is an important research direction in the field of computer vision. In this work, we incorporate the skeleton on the Grassmann manifold in order to model the human action as a trajectory. Given the couple of matched points on the Grassmann manifold, we introduce the special orthogonal group SO(3) to exploit the rotation ignored by the Grassmann manifold. In fact, our objective is to define the best weighted linear combination between distances in Grassmann and SO(3) man-ifolds according to the nature of action, while modeling human actions by temporal trajectories and finding the best weighted combination. The effectiveness of combining the two non-Euclidean spaces was validated on three standard challenging 3D human action recognition datasets (G3D-Gaming, UTD-MHAD multimodal action and Florence3D-Action), and the preliminary results confirm the accuracy of the proposed method comparatively to relevant methods from the state of the art.