Essay On Feature Extraction

• Feature Extraction : this is the most important stage for automated markerless capture systems whether for gait recognition, activity classification or other

In 2010, S.Arivazhagan, R.Newlin Shebiah, S.Selva Nidhyanandhan, L.Ganesan had done their research, and they published their research named ‘Fruit Recognition using Color and Texture Features’. The objective of this project is to recognize a fruit based on four common features including intensity, color, shape, and texture of the object. Moreover,

These 16 features included 12 features calculated based on the 6 multispectral bands, which is mean value and standard deviation of these bands. In addition, we chose intensity, texture-variance, texture-mean, and NDVI (Normalized Difference Vegetation Index) for classifications. Finally, training samples were selected for each classification category based on the previously segmented and merged objects

Thus our proposed optimal feature subset selection based on multi-level feature subset selection produced better results based on number of subset feature produced and classifier performance. The future scope of the work is to use these features to annotate the image regions, so that the image retrieval system can retrieve relevant images based on image semantics.

Feature extraction is most important step in the process of CBIR. Feature extraction is a method of transforming input data into set of features [2]. The different feature extraction is colour, texture and shape. Features are classified as low level and high level. The low level feature contains colour, texture and high level feature contains shape. The various feature extraction are describing in below:

The particular finding parameters (i.e. threshold, fwhm, sharpness, and roundness) were tuned to the data to maximize true identification yields as determined by visual inspection. Some parameters like fwhm and signal threshold were set adaptively by using image based diagnostics like a first order estimate of the Gaussian fit PSF and the background noise levels, respectively. It was particularly tricky to find parameters that worked well for both clear and highly nebulous regions with small features. Also, the parameters had to be tuned for particular filters and co-adding depths. The subsequent photometry routines for grouping, nstar fitting, and PSF subtraction were all part of the Photoutils package. A sigma clipped median was used to estimate the

Paper [1] presented Integrated Multiple Features for Tumor Image Retrieval Using Classifier and Feedback Methods. This paper presents an effective approach in which the region of the object is extracted with the help of multiple features ignoring the background of the object by employing edge following segmentation method followed by extracting texture and shape characteristics of the images. The former is extracted with the help of Steerable filter at different orientations and radial Chebyshev moments are used for extracting the later.

Cordelli et al. [11] consider a heterogeneous set of texture features belonging to different categories, statistical descriptors, spectral measures, local binary pattern (LBP) and morphological descriptors.

Feature selection (FS) methods have been used in the since 70s, using in the fields of statistics and pattern recognition. Pattern recognition system is one of the most important and indispensable tasks in overcome the curse of dimensionality problem, which forms a motivation for using a suitable feature selection method. According to their working principles, there are two types of methods are using in feature selection: methods which select the best subset of features that has a certain number of features And methods which select the best subset of features according to their own principles, independent of outside size measures [base].

Feature selection is a method used for reducing number of dimensions of a dataset by removing irrelevant and redundant attributes. Given a set of attributes F and a target class C, goal of feature selection is to find a minimum set of F that will yield highest accuracy (for C) for the classification task. Although

There are several ways to select the best features. \cite{Thomas}. Also, it has been shown that selection of the number of the features for classification, neighbors and the predictors are very deterministic in the quality of the classification \cite{NLi}.

Tested for a number of datasets and it is found that accuracy of image recognition for proposed approach is higher. The results are evaluated for different resolutions of the image and it is found that even for low resolution accuracy is considerably more for proposed approach.

Face recognition is a technology that is capable of identifying a person from a multimedia object such as a

Image classification makes use of analyzing various properties of various image features and then sets the data into classes. Classification methods typically uses two levels for the process of classification: training and testing. Primarily in the training phase, characteristics of the image features are extracted and, based on these, a unique description of each classification category, i.e. training class, is created\cite{mohri2012foundations}. In the subsequent testing phase, these feature-space partitions are used to classify image features.\\

Image Processing is a technique to enhance raw images received from cameras/sensors placed on satellites, space probes and aircrafts or pictures taken in normal day-today life for various applications. Various techniques have been developed in Image Processing during the last four to five decades. Most of the techniques are developed for enhancing images obtained from unmanned spacecrafts, space probes and military reconnaissance flights. Image Processing systems are becoming popular due to easy availability of powerful personnel computers, large size memory devices, graphics software’s etc. The common steps in image processing are image scanning, storing, enhancing and interpretation.