Divi Galih Prasetyo Putri, Daniel Siahaan · 2016
Evolution and maintenance processes are important but time consuming and expensive. It is very important to make the processes effective and efficient. A software developer can use resource like user opinion data to get information, such as user request, bug report, and user experience. It represents user needs and can be used to help allocate the necessary effort of software evolution and maintenance. The amount of user opinion data is very large and is difficult manually process them. A Recent study has tried to implement collocation finding method to extract software features from user opinion data. However, it is not able to extract non-frequently mentioned features. In this paper, we proposed an improvement for software feature extraction from user opinion data. Linguistic rules were used to complement collocation finding method. Feature pruning was also added to eliminate irrelevant features. The result shows that the proposed method is able to extract more features than collocation finding method.
Cited by 9
Wenbo Cao, Robert Haralick, Affine Transformation · 2008
Fisher’s discriminant analysis Fukunaga–Koontz transformation Kullback–Leibler divergence a b s t r a c t Dimension reduction methods are often applied in machine learning and data mining problems. Linear subspace methods are the commonly used ones, such as principal component analysis (PCA), Fisher’s linear discriminant analysis (FDA), common spatial pattern (CSP), et al. In this paper, we describe a novel feature extraction method for binary classification problems. Instead of finding linear subspaces, our method finds lower-dimensional affine subspaces satisfying a generalization of the Fukunaga– Koontz transformation (FKT). The proposed method has a closed-form solution and thus can be solved very efficiently. Under normality assumption, our method can be seen as finding an optimal truncated spectrum of the Kullback–Leibler divergence. Also we show that FDA and CSP are special cases of our proposed method under normality assumption. Experiments on simulated data show that our method performs better than PCA and FDA on data that is distributed on two cylinders, even one within the other. We also show that, on several real data sets, our method provides statistically significant improvement on test set accuracy over FDA, CSP and FKT. Therefore the proposed method can be used as another preliminary data-exploring tool to help solve machine learning and data mining problems. & 2008 Elsevier Ltd. All rights reserved. 1.
Cited by 0
David Tschumperlé, Christophe Tilmant, Vincent Barra · 2023
Cited by 0
2014
This work is focused towards determining important contents for the description of a video stream and addresses the problem of generating smooth, coherent natural language description of videos in time series. NLDs of visual scenes are needed as they are more human friendly and capture relationships among keywords; clarifying the context related to separate keywords.
Cited by 0
G.J. Awcock, R. Thomas · 1995
The feature extraction aspect of image analysis seeks to identify inherent characteristics, or features, of objects found within an image. These characteristics are used to describe the object, or attributes of the object, prior to the subsequent task of classification. Feature extraction operates on two-dimensional image arrays but produces a list of descriptions, or a ‘feature vector’ (note the change in information format indicated on the generic model, section 1.4.1).
Cited by 0
2013
Cited by 0
Reda Alhajj, Jon Rokne · 2018
Cited by 0
2015
Cited by 0
Yide Ma, Kun Zhan, Zhaobin Wang · 2010
Cited by 2
2013
Cited by 0
Paolo Remagnino, Simon Joseph Mayo, Paul Wilkin et al. · 2016
Cited by 1
James Eric Mason, Issa Traoré, Isaac Woungang · 2016
Cited by 0
2010
Cited by 0Open access
K. Hanke A, W. Boehler B · 2014
The Hofkirche in Innsbruck, Austria, with its tomb of German Emperor Maximilian I is one of the most famous and outstanding historical monuments in Central Europe. For centuries the Cenotaph (i.e. empty tomb) was separated from the visitors by a black iron lattice. In addition, the fine caved marble plates were covered by glass. Because of a basic restoration of the tomb, lattice and glass plates were removed for the first time ever since its construction in the 16 th century. For a short period in May 2002 all sides were accessible after the temporary housing of the restoration technicians had been removed from one side and not yet been moved to the other side for the second restoration period. This could be used for a complete metric documentation of the object. Both, close-range photogrammetry and 3D scanning techniques were used. Photogrammetric images consisted of stereo pairs and separate color images. 3D scanning was accomplished with a MENSI S25 laser scanner for the overall structures and a GOM ATOS II structured light scanner at high resolution for the relief plates. Fig. 1: Overview model of the cenotaph derived from scanned data * Corresponding author. Different methods can be used to visualize the meshed surface model. Line plots from the photogrammetric stereo models do not really give an adequate representation of this object. 3D visualization using the scanning results can achieve a much better impression of the complicated geometry after data modeling, error correction, and filling in remaining ‘holes’. In order to model the complex geometry, it is necessary to use huge amounts of data. The model in its highest resolution consists of more than 1.000.000.000 triangles. Because of restrictions in the hard- and software presently available, the high resolution model has to be
Cited by 0
Anindita Das Bhattacharjee · Advances in multimedia and interactive technologies book series · 2016
Accessibility problem is relevant for audiovisual information, where enormous data has to be explored and processed. Most of the solutions for this specific type of problems point towards a regular need of extracting applicable information features for a given content domain. And feature extraction process deals with two complicated tasks first deciding and then extracting. There are certain properties expected from good features-Repeatability, Distinctiveness, Locality, Quantity, Accuracy, Efficiency, and Invariance. Different feature extraction techniques are described. The chapter concentrates of taking a survey on the topic of Feature extraction and Image formation. Here both image and video are considered to have their feature extracted. In machine learning, pattern recognition and in image processing has significant contribution. The feature extraction is one of the common mechanisms involved in these two techniques. Extracting feature initiates from an initial data set of measured data and constructs derived informative values which are non redundant in nature.
Cited by 0
Claudia Fuchs, Stephan Heuel · 1998
Contents 1 Introduction 3 1.1 General Problems with Feature Extraction . . . . . . . . . . . . . . . . . . 4 2 Overview on Existing Techniques 5 2.1 Extraction of Points . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6 2.1.1 General Principles for Point Extraction . . . . . . . . . . . . . . . . 6 2.1.2 Point Templates . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7 2.1.3 Corner Detection by Curvature . . . . . . . . . . . . . . . . . . . . 8 2.1.4 Point Detection by Optimization . . . . . . . . . . . . . . . . . . . 8 2.2 Extraction of Edges . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9 2.2.1 General Principles for Edge Extraction . . . . . . . . . . . . . . . . 9 2.2.2 Edge Regions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12 2.2.3 Edge Pixels . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13 2.2.4 Edge Elements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14 2
Cited by 0
Ziqian Wan, Alex J. Nelson, Tao Li et al. · 2012
Geolocation data have high value to forensic investigators because computer activities may be associated with physical locations in the past. However, locating and extracting useful location information from an off-line disk image is a difficult problem. Most forensic investigations employ tools that focus on extracting content, such as emails, databases, and hidden or deleted data, and then manually investigate the results with practices like keyword searches. While this can work on a drive-by-drive basis, without a uniform approach to the location question, it is easy for an investigator to miss an answer that could be found from an evaluated technique known to other investigators. To determine drive location, we develop a two-step approach that analyzes a drive image for geolocation purposes, finding substantial location information in HTTP headers from common and default sources. First, we extract HTTP headers from the memory page (swap) files that reside on the hard drive. Second, we apply a weight based algorithm that parses those headers to determine the past geographical locations of the drive. We apply our method to drive images from the publicly available M57 Patents corpus and identify the hard drives ’ location with low recall but high precision. 1
Cited by 0
2011
Cited by 0
Soheil Keshmiri (6338894), Maryam Alimardani (3106236), Masahiro Shiomi (5663563) et al. · Figshare · 2020
<p>The function H calculates the entropy of its input time series as per <a href="http://www.plosone.org/article/info:doi/10.1371/journal.pone.0230853#pone.0230853.e001" target="_blank">Eq (1)</a>. <i>X</i><sup><i>θ</i></sup>, <i>X</i><sup><i>α</i></sup>, and <i>X</i><sup><i>β</i></sup> are the frequency components associated with EEG time series X. We used the resulting feature vector <i>V</i> = [<i>H</i>(<i>X</i><sup><i>θ</i></sup>), <i>H</i>(<i>X</i><sup><i>α</i></sup>), <i>H</i>(<i>X</i><sup><i>β</i></sup>)] in our analyses.</p>
Cited by 0Open access
Pham, Lam · Zenodo (CERN European Organization for Nuclear Research) · 2021
Feature extraction
Cited by 0Open access