Suitable Set Of Features Research Articles

AUTOMATIC MAPPING OF GLACIER BASED ON SAR IMAGERY BY BENEFITS OF FREELY OPTICAL AND THERMAL DATA

Abstract. For many research applications like water resources evaluation, determination of glacier specific changes, and for calculation of the past and future contribution of glaciers to sea-level change, parameters about the size and spatial distribution of glaciers is crucial. In this paper, an automatic method for determination of glacier surface area using single track high resolution TerraSAR-X imagery by benefits of low resolution optical and thermal data is presented. Based on the normalized difference snow index (NDSI) and land surface temperature (LST) map generated from optical and thermal data combined with a surface slope data, a low resolution binary mask was derived used for the supervised classification of glacier using SAR imagery. Then, a set of suitable features is derived from the SAR intensity image, such as the texture information generated based on the gray level co-occurrence matrix (GLCM), and the intensity values. With these features, the glacier surface is discriminated from the background by Random Forests (RF) method.

Evaluation of Shape and Color Features for Classification of Four Paddy Varieties

This research is aimed at evaluating the shape and color features using the most commonly used neural network architectures for cereal grain classification. An evaluation of the classification accuracy of shape and color features and neural network was done to classify four Paddy (Rice) grains, viz. Karjat-6, Ratnagiri-2, Ratnagiri-4 and Ratnagiri-24. Algorithms were written to extract the features from the high-resolution images of kernels of four grain types and use them as input features for classification. Different feature models were tested for their ability to classify these cereal grains. Effect of using different parameters on the accuracy of classification was studied. The most suitable feature set from the features was identified for accurate classification. The Shape-nColor feature set outperformed in almost all the instances of classification.

A novel artificial neural networks assisted segmentation algorithm for discriminating almond nut and shell from background and shadow

Segmentation is one of the main steps in image processing, as it influences the accuracy of other processes such as feature selection and classification. In this study, an effective method based on a combined image processing and machine learning was presented and evaluated for segmenting almond images with different classes such as normal almond, broken and split almond, shell of almond, wrinkled almond and doubles or twins almond. One of the major difficulties encountered in segmenting almonds was the existence of shadow on the background of the acquired images. Another difficulty was separating almonds with various shapes and colors from input images. To implement an effective algorithm, initially a suitable set of color features was extracted from the images. Then, sensitivity analysis was used to select the best features. Finally, artificial neural networks (ANNs) were adopted to classify the images into three categories, namely, object, shadow and background. The optimum ANN classifier had a 8-5-3 structure, i.e., it was consisted of an input layer with eight input variables, one hidden layer with five neurons and three neurons as output. To evaluate the performance of the proposed method, the results of our optimum ANN model were compared with Otsu, dynamic thresholding and watershed methods. The mean values of sensitivity, specificity and accuracy for object class (detected almonds from images) achieved by using the proposed method were 96.88, 99.21 and 98.82, respectively. It gave a better accuracy than the mentioned methods. In addition, the proposed method was able to separate the almonds from the background and shadows more efficiently. The processing time of the proposed method was 1.35s which makes it possible for real time applications.

A new thermographic NDT for condition monitoring of electrical components using ANN with confidence level analysis

Infrared thermography technology is one of the most effective non-destructive testing techniques for predictive faults diagnosis of electrical components. Faults in electrical system show overheating of components which is a common indicator of poor connection, overloading, load imbalance or any defect. Thermographic inspection is employed for finding such heat related problems before eventual failure of the system. However, an automatic diagnostic system based on artificial neural network reduces operating time, human efforts and also increases the reliability of system. In the present study, statistical features and artificial neural network (ANN) with confidence level analysis are utilized for inspection of electrical components and their thermal conditions are classified into two classes namely normal and overheated. All the features extracted from images do not produce good performance. Features having low performance reduce the diagnostic performance. The study reveals the performance of each feature individually for selecting the suitable feature set. In order to find the individual feature performance, each feature of thermal image was used as input for neural network and the classification of condition types were used as output target. The multilayered perceptron network using Levenberg–Marquardt training algorithm was used as classifier. The performances were determined in terms of percentage of accuracy, specificity, sensitivity, false positive and false negative. After selecting the suitable features, the study introduces the intelligent diagnosis system using suitable features as inputs of neural network. Finally, confidence percentage and confidence level were used to find out the strength of the network outputs for condition monitoring. The experimental result shows that multilayered perceptron network produced 79.4% of testing accuracy with 43.60%, 12.60%, 21.40, 9.20% and 13.40% highest, high, moderate, low and lowest confidence level respectively.

Online Direction Classification and Recognition Using an Intuitive Tactile Communication

This paper proposed an online direction classifying method for constructing an intuitive tactile communication during human-robot cooperation. The proposed approach abstracts a suitable feature set from a tactile array sensor equipped on a hand-bar. This lower computation feature extraction method analyze the weighting values concerned with oriental information from principle component analysis (PCA), together with support vector machines (SVM) classifier for direction classification and recognition. Experimental results showed an average accuracy of 96.3% and a low costs of 512μs with respect to different handle gestures of the 6 touch directions, which is practicable utilized for human-robot cooperation based on tactile recognition.

Motion Tracking, Retrieval and 3D Reconstruction from Video

The work at hand presents a novel data-driven framework for 3D full body human motion reconstruction from uncalibrated monocular video data. To this end, we develop a knowledge base by taking 2D samples of the motion capture library from dierent viewing directions. This allows later steps to handle 2D query videos without any information on the viewing direction. We detect and track features from input video sequences by utilizing low-level image based feature detection techniques like MSER and SURF. This process is stabilized by back projection of high-level 3D prior information obtained from the motion capture library to the image plane. Extraction of suitable feature sets from both, input control signals and motion capture data, enables us to retrieve the best relevant prior poses from the motion capture library by employing fast motion retrieval techniques. Finally, 3D motion sequences are reconstructed by non-linear energy minimization, that takes into account multiple prior terms. Furthermore, we propose a method to estimate camera parameters from input video itself and sampling of motion capture library.

Modified CC-LR algorithm with three diverse feature sets for motor imagery tasks classification in EEG based brain–computer interface

Motor imagery (MI) tasks classification provides an important basis for designing brain–computer interface (BCI) systems. If the MI tasks are reliably distinguished through identifying typical patterns in electroencephalography (EEG) data, a motor disabled people could communicate with a device by composing sequences of these mental states. In our earlier study, we developed a cross-correlation based logistic regression (CC-LR) algorithm for the classification of MI tasks for BCI applications, but its performance was not satisfactory. This study develops a modified version of the CC-LR algorithm exploring a suitable feature set that can improve the performance. The modified CC-LR algorithm uses the C3 electrode channel (in the international 10–20 system) as a reference channel for the cross-correlation (CC) technique and applies three diverse feature sets separately, as the input to the logistic regression (LR) classifier. The present algorithm investigates which feature set is the best to characterize the distribution of MI tasks based EEG data. This study also provides an insight into how to select a reference channel for the CC technique with EEG signals considering the anatomical structure of the human brain. The proposed algorithm is compared with eight of the most recently reported well-known methods including the BCI III Winner algorithm. The findings of this study indicate that the modified CC-LR algorithm has potential to improve the identification performance of MI tasks in BCI systems. The results demonstrate that the proposed technique provides a classification improvement over the existing methods tested.

Reliability estimation for component-based software product lines

The objective of the software product line engineering paradigm is to enhance the largescale reuse of common core assets within a target domain. Reuse is facilitated by systematically organizing and modeling the core assets and the relationships between them. One of the main core assets of a domain is the model for representing the available functional aspects, often known as features, within structured forms such as feature models. The selection and composition of the most suitable or desirable set of features for a given purpose allows the rapid development of new final products from the software product line. Product developers are, in most cases, not only interested in building applications that possess certain functional characteristics but are also concerned with nonfunctional properties of the final product, such as reliability. To this end, we propose a component based software product line reliability estimation model that is able to provide lower and upper reliability bounds guarantees for a software product line feature model, its specializations and configurations. Our model builds on top of the reliability of the individual features that are present in the product line and provides best- and worst-case estimates. Our work is based on an essential and widely used assumption that features are implemented using self-contained software components or services whose reliability can be determined independently. We also propose reliability-aware configuration methods that ensure the satisfaction of both functional and reliability requirements during the application development process. We offer our observations and insight into the performance of our reliability estimation model and provide analysis of its advantages and shortcomings.

Sensor-oriented feature usability evaluation in fingerprint segmentation

Existing fingerprint segmentation methods usually process fingerprint images captured by different sensors with the same feature or feature set. We propose to improve the fingerprint segmentation result in view of an important fact that images from different sensors have different characteristics for segmentation. Feature usability evaluation, which means to evaluate the usability of features to find the personalized feature or feature set for different sensors to improve the performance of segmentation. The need for feature usability evaluation for fingerprint segmentation is raised and analyzed as a new issue. To address this issue, we present a decision-tree-based feature-usability evaluation method, which utilizes a C4.5 decision tree algorithm to evaluate and pick the best suitable feature or feature set for fingerprint segmentation from a typical candidate feature set. We apply the novel method on the FVC2002 database of fingerprint images, which are acquired by four different respective sensors and technologies. Experimental results show that the accuracy of segmentation is improved, and time consumption for feature extraction is dramatically reduced with selected feature(s).

DETERMINATION OF GLACIER SURFACE AREA USING SPACEBORNE SAR IMAGERY

Abstract. Glaciers are very important climate indicators. Although visible remote sensing techniques can be used to extract glacier variations effectively and accurately, the necessary data are depending on good weather conditions. In this paper, a method for determination of glacier surface area using multi-temporal and multi-angle high resolution TerraSAR-X data sets is presented. We reduce the "data holes" in the SAR scenes affected by radar shadowing and specular backscattering of smooth ice surfaces by combining the two complementary different imaging geometries (from ascending and descending satellite tracks). Then, a set of suitable features is derived from the intensity image, the texture information generated based on the gray level co-occurrence matrix (GLCM), glacier velocity estimated by speckle tracking, and the interferometric coherence map. Furthermore, the features are selected by 10-foldcross- validation based on the feature relevance importance on classification accuracy using a Random Forests (RF) classifier. With these most relevant features, the glacier surface is discriminated from the background by RF classification in order to calculate the corresponding surface area.

Keyword spotting using an evolutionary-based classifier and discriminative features

Keyword spotting refers to detection of all occurrences of any given keyword in input speech utterances. In this paper, we define a keyword spotter as a binary classifier that separates a class of sentences containing a target keyword from a class of sentences which do not include the target keyword. In order to discriminate the mentioned classes, an efficient classification method and a suitable feature set are to be studied. For the classification method, we propose an evolutionary algorithm to train the separating hyper-plane between the two classes. As our discriminative feature set, we propose two confidence measure functions. The first confidence measure function computes the possibility of phonemes presence in the speech frames, and the second one determines the duration of each phoneme. We define these functions based on the acoustic, spectral and statistical features of speech. The results on TIMIT indicate that the proposed evolutionary-based discriminative keyword spotter has lower computational complexity and higher speed in both test and train phases, in comparison to the SVM-based discriminative keyword spotter. Additionally, the proposed system is robust in noisy conditions.

Class Specific Feature Selection for Identity Validation using Dynamic Signatures

Classification of the biometrics data for identity validation can be modeled as a single-class problem, where the identity is confirmed by comparing the biometrics of the unknown person with those of the claimed identity. However, current feature selection techniques do not differentiate between single-class and multi-class problems when determining the suitable feature set and select the feature-set that is suitable for representing or discriminating for all the available classes. This may not be the best representation of the biometrics data of an individual because different people may have differences in the most suitable features to represent their biometrical data. In this paper, a class-specific feature selection method has been proposed and experimentally validated using dynamic signatures. This method is based on the coefficient of variance within the feature set, where the features with smaller variance are selected and the ones with larger variance are rejected. The proposed technique was compared with the other feature selection methods, and the results show that a significant improvement in the classification accuracy, specificity and sensitivity was obtained when using class-specific feature selection.

An eye–hand data fusion framework for pervasive sensing of surgical activities

This paper describes a generic framework for activity recognition based on temporal signals acquired from multiple input modalities and demonstrates its use for eye–hand data fusion. As a part of the data fusion framework, we present a multi-objective Bayesian Framework for Feature Selection with a pruned-tree search algorithm for finding the optimal feature set(s) in a computationally efficient manner. Experiments on endoscopic surgical episode recognition are used to investigate the potential of using eye-tracking for pervasive monitoring of surgical operation and to demonstrate how additional information induced by hand motion can further enhance the recognition accuracy. With the proposed multi-objective BFFS algorithm, suitable feature sets both in terms of feature relevancy and redundancy can be identified with a minimal number of instruments being tracked.

A Study on the Analysis of Genetic Algorithms with Various Classification Techniques for Feature Selection

world of curse of dimensionality feature selection plays a very important role in reducing the entire feature collection with the limited subset of features. Reducing the number of features pave way for various advantaged as well as simplifies the task. Feature selection means finding the suitable set of features which will contribute the most of it to the solution with minimal or null error rate. Selected features are to be tested with the help of classifiers, so that the subset of selected features can be proved to be optimal when compared to other features subsets individually as well as a group. Genetic algorithms are now days play a vital role among any other methodology in selecting the features based on the Theory of Evolution and on the Survival of the fitness. It is a heuristic approach. To cooperate with the GA approach we have the classifiers which will go hand in hand to bring out the final set of features along with their prediction accuracy. In this paper I have analyzed four of the classifiers and compared them with their performance and the unit of accuracy.

Splice sites prediction of Human genome using length-variable Markov model and feature selection

As the rapid increase of DNA sequences, there is a crucial need of effective methods to detect genes and genes’ structures, among which splice sites prediction plays a key role. There are conservative segments on the junctions between introns and exons, which can help us predict splice sites by computational methods; however it is unclear that which nucleotides contribute to the splicing process, so it is necessary to select a suitable set of features to accomplish the prediction of splice sites. A length-variable Markov model is proposed in this paper. By the length-variable model, a suitable subset of features can be chosen as the detecting features for each candidate splice site according to the ratio of likelihood at each position. The results of our experiments show that our models not only achieve higher prediction accuracy than the basic Markov model and some present methods, but also preserve the feature of low time cost as the basic Markov model does.

Features for image retrieval: an experimental comparison

An experimental comparison of a large number of different image descriptors for content-based image retrieval is presented. Many of the papers describing new techniques and descriptors for content-based image retrieval describe their newly proposed methods as most appropriate without giving an in-depth comparison with all methods that were proposed earlier. In this paper, we first give an overview of a large variety of features for content-based image retrieval and compare them quantitatively on four different tasks: stock photo retrieval, personal photo collection retrieval, building retrieval, and medical image retrieval. For the experiments, five different, publicly available image databases are used and the retrieval performance of the features is analyzed in detail. This allows for a direct comparison of all features considered in this work and furthermore will allow a comparison of newly proposed features to these in the future. Additionally, the correlation of the features is analyzed, which opens the way for a simple and intuitive method to find an initial set of suitable features for a new task. The article concludes with recommendations which features perform well for what type of data. Interestingly, the often used, but very simple, color histogram performs well in the comparison and thus can be recommended as a simple baseline for many applications.

Optimizing plane-to-plane positioning tasks by image-based visual servoing and structured light

This paper considers the problem of positioning an eye-in-hand system so that it becomes parallel to a planar object. Our approach to this problem is based on linking to the camera a structured light emitter designed to produce a suitable set of visual features. The aim of using structured light is not only for simplifying the image processing and allowing low-textured objects to be considered, but also for producing a control scheme with nice properties like decoupling, convergence, and adequate camera trajectory. This paper focuses on an image-based approach that achieves decoupling in all the workspace, and for which the global convergence is ensured in perfect conditions. The behavior of the image-based approach is shown to be partially equivalent to a 3-D visual servoing scheme, but with a better robustness with respect to image noise. Concerning the robustness of the approach against calibration errors, it is demonstrated both analytically and experimentally

What should be expected from feature selection in small-sample settings

High-throughput technologies for rapid measurement of vast numbers of biological variables offer the potential for highly discriminatory diagnosis and prognosis; however, high dimensionality together with small samples creates the need for feature selection, while at the same time making feature-selection algorithms less reliable. Feature selection must typically be carried out from among thousands of gene-expression features and in the context of a small sample (small number of microarrays). Two basic questions arise: (1) Can one expect feature selection to yield a feature set whose error is close to that of an optimal feature set? (2) If a good feature set is not found, should it be expected that good feature sets do not exist? The two questions translate quantitatively into questions concerning conditional expectation. (1) Given the error of an optimal feature set, what is the conditionally expected error of the selected feature set? (2) Given the error of the selected feature set, what is the conditionally expected error of the optimal feature set? We address these questions using three classification rules (linear discriminant analysis, linear support vector machine and k-nearest-neighbor classification) and feature selection via sequential floating forward search and the t-test. We consider three feature-label models and patient data from a study concerning survival prognosis for breast cancer. With regard to the two focus questions, there is similarity across all experiments: (1) One cannot expect to find a feature set whose error is close to optimal, and (2) the inability to find a good feature set should not lead to the conclusion that good feature sets do not exist. In practice, the latter conclusion may be more immediately relevant, since when faced with the common occurrence that a feature set discovered from the data does not give satisfactory results, the experimenter can draw no conclusions regarding the existence or nonexistence of suitable feature sets. http://ee.tamu.edu/~edward/feature_regression/

ASSESSMENT OF NEURAL NETWORK SCHEMES TO CLASSIFY CLOUD DATA

Using remotely-sensed data from the Tropical Rainfall Measuring Mission (TRMM), a cloud classification study was undertaken employing neural network schemes. The objective of this study was to assess the accuracy of each scheme for classifying clouds. In the first phase, a data preprocessing and feature selection scheme was undertaken to identify a suitable set of features that could be useful in cloud classification. In the next phase, seven neural network classification schemes were implemented to understand the utility of each of these schemes. Parametric schemes performed poorly, while the perceptron, K-nearest neighbor approaches and the least means square algorithm yielded promising results. Further study is proposed so as to improve rainfall prediction.

A PHYLOGENETIC CLASSIFICATION OF THE HOMOSPOROUS FERNS

More than probably any other group of plants, the pteridophytes, particularly the group of ferns, have long remained problematical from the point of view of taxonomy. This is scarcely surprising, as in this group the search is still continuing for a suitable set of catholic morphological features which are amenable to precise description, and which indicate the possible trends of evolution within the group. Also, the vast majority of these plants (nearly 90% of the number of known species) are strictly tropical with no temperate representatives, and these were morphologically almost unknown until recently. However, the revival of interest in the morphology of these plants in the present century has brought to light a wealth of information about them. The monumental morphological investigations undertaken by Prof. F. O. Bower early in the century proved beyond doubt that the taxonomic criteria used by earlier workers were inadequate to express phylogeny. The modern concept that taxonomy should also express phylogeny led to a reconsideration of the classification of these plants, and during the past three decades no less than seven systems came to be proposed (Christensen, 1938; Ching, 1940; Holttum, 1947; Copeland, 1947; Reimers, 1954; Alston, 1956; Pichi-Sermolli, 1958). All these differ from each other in quite substantial features, mainly due to the differences in the relative significance attributed to different morphological features by the various authors; the addition of fresh morphological data has also played a significant role in interpreting relationships of many groups. Of late, there have been significant additions to our knowledge relating to several morphological criteria of the pteridophytes. Among these are the chromosome counts and details of prothallial morphology, spore morphology and anatomy, which are now known for a vast number of genera. An analysis of these and other morphological criteria suggests that the current interpretations of taxonomy and evolution of the homosporous ferns are possibly not totally acceptable. A new scheme for the classification of these plants is therefore presented here: