Earticle

This paper aims at making a contribution to a typical recognition scheme by means of providing position information of the recognized objects. In the near future, robotics systems are expected to be able to provide services in humans’daily life which can only be achieved if they are designed with advanced intelligent vision systems. Highly motivated from that fact, this work presents a framework capable of recognizing objects and estimating their location in the 3D space. The method excels in simplicity and computational cost, whilst its database can be easily adopted to the needs of simultaneous multi-object recognition. In real-life scenarios, lighting conditions may alter drastically and, along with possible shadow effects, may affect directly the efficiency of the visual data encoding scheme. Towards this end, recently proposed image enhancement methods emphasize in reducing the effect of unfavorable illumination conditions and their consequences. The proposed 3D position estimation framework was assessed under several image-enhancement conditions by means of selecting the most appropriate pre-processing algorithm.

Palmprint recognition is a biometric method to automatically identify a person’s identity. In this paper, phase congruency method is proposed to extract features from a palm-print image for authentication. The phase congruency is one of the promising methods to analyse the image as it is invariant to image contrast and therefore can extract reliable features under varying illumination conditions. In this paper, the phase congruency features in 6 different orientations are arranged in such a manner to get the best combination of orientations for authentication. The hand image is pre-processed to get the desired Region of Interest (ROI) / palmprint. The palmprint features are extracted by phase congruency method and are stored in feature vector. The 20 different types of feature vectors are prepared using different combination of orientations. Hamming distance similarity measurement with Sliding window is used to compare the similarity/dissimilarity between two feature vectors. Experiments were developed on a database of 600 images from 100 individuals, with five image samples per individual for training and one image sample per individual for testing. The accuracy of 97.3% can be achieved using FV11.

Understanding the polarization of light is becoming increasingly important in the study of non linear optics in computer vision. This is caused by the polarization state of light can provides an essential information concerning the observed surface orientations more proficient than the intensity information from conventional imaging system in many light condition. That is the reason why polarization imaging system can be widely uses for many application in computer vision field, such as object recognition, shape estimation or object segmentation. In the other side, stereo vision infers scene geometry from images pair with different viewpoints. Using Stereo vision can improve image understanding technique by obtaining depth information from pairs of digital images. Partially, many researchers have been done a lot of method in both imaging vision technique. However, there is very little research to combine the stereo vision with polarization imaging. The motivation of proposed work is come from nature, from many animals that use two eyes to sense a polarization light as a further source of visual information. The developed prototype is made of a stereo camera set-up with two liquid crystal components in front of the lenses. This article also describes the geometric and the photometric calibration process that is required and provides algorithms that enable to extract both three-dimensional information and polarization information.

In image processing wavelet transform methods are known to provide efficient schemes for detecting edges and reducing noise. However, they are unable to differentiate between noise and small details and therefore remove the small details resulting in the formation of oscillations around edges. Methods based on isotropic nonlinear equations have been used to enhance edges in images and to reduce oscillations around the edges but are not capable of removing noise effectively around them. In this paper, we solve the isotropic nonlinear equation in a geometric transform space and thereby reducing both noise and oscillations around the edges. The proposed synergic method combines the advantages of both techniques. The numerical experiments involve the investigation of two-dimensional geometric transforms in the detection of edges in noisy images.

Image compression is a widely addressed researched area. Many compression standards are in place. But still here there is a scope for high compression with quality reconstruction. The JPEG standard makes use of Discrete Cosine Transform (DCT) for compression. The introduction of the wavelets gave a different dimensions to the compression. This paper aims at the analysis of compression using DCT and Wavelet transform by selecting proper threshold method, better result for PSNR have been obtained. Extensive experimentation has been carried out to arrive at the conclusion.

Classifying cork stopper into group required large set of visual features. Selecting an optimal feature subset from large input feature set speeds up classification task and improve the classifier accuracy. Traditional feature selection methods, such as sequential forward selection, sequential backward selection, and sequential forward floating search are costly to implement. This paper we propose a feature selection method known as principal feature analysis that exploits the structure of the principal components of a feature set to find a subset of the original features information and support vector machines (SVMs) for classification. The experimental result show that the proposed method for SVM based classifier is lot faster than PCA and ICA based methods. It is also leads to better performance when the same number of principal/independent components is used and consistently picks the best subset of features in terms of sum-squared-error compared to competing methods.

The paper will present a novel approach for solving face recognition problem. Our method combines 2D Principal Component Analysis (2DPCA), one of the prominent methods for extracting feature vectors, and Support Vector Machine (SVM), the most powerful discriminative method for classification. Experiments based on proposed method have been conducted on two public data sets FERET and AT&T; the results show that the proposed method could improve the classification rates.

The paper addresses a new QRS complex geometrical feature extraction technique as well as its application for supervised electrocardiogram (ECG) heart-beat type classification. Toward this objective, after detection and delineation of major events of the ECG signal via an appropriate algorithm, each QRS region and also its corresponding discrete wavelet transform (DWT) are supposed as virtual images and each of them is divided into eight polar sectors. Then, the curve length of each excerpted segment is calculated and is used as the element of the feature space. Afterwards, an appropriate fuzzy network classifier aimed for recognizing several heart-beat types is preliminarily designed. To propose a new classification strategy with adequate robustness against noise, artifacts and arrhythmic outliers, the fuzzy rules parameterization and determination stages were fulfilled using the fuzzy c-means (FCM) and the subtractive clustering techniques. To show merit of the new proposed algorithm, it was applied to 4 number of arrhythmias (Normal, Left Bundle Branch Block-LBBB, Right Bundle Branch Block-RBBB, Paced Beat-PB) belonging to 12 records of the MIT-BIH Arrhythmia Database and the average accuracy values Acc=94.58% and Acc=97.41% were achieved for FCM-based and subtractive clustering-based fuzzy-logic classifiers, respectively. To evaluate operating characteristics of the new proposed fuzzy classifier, the obtained results were compared with similar peer-reviewed studies in this area.

Today, the primary constrain in wireless communication system is limited bandwidth and power. Wireless systems involved in transmission of speech envisage that efficient and effective methods be developed (bandwidth usage & power) to transmit and receive the same while maintaining quality-of-speech, especially at the receiving end. Speech coding is a technique, since the era of digitization (digital) and computerization (computational and processing horsepower - DSP) that has been a material- of- research for quite sometime amongst the scientific and academic community. Amongst all elements of the communication system (transmitter, channel and receiver), transmission channel (carrier of information/data, also called the medium) is the most critical and plays a key role in the transmission and reception of information/data. This paper proposes some modifications in the Long Term Gain and selection of Grid position sections of 13kbps GSM Full Rate coder so that it can be mapped to bitrates of GSM Enhanced Full Rate at12.2kbps. Here, Objective analysis is carried out on a proposed system using MATLAB to evaluate its performance and the results obtained are then compared to the results of GSM Full Rate coder. Proposed coder, at bitrates of 12.2 kbps, provides moderately low computational complexity [1] as its implementation is with reference to GSM full rate coder rather than using Algebraic Code Excited Linear Prediction (ACELP) algorithm in case of standard GSM EFR coder but it offers small degradation in speech quality compared to standard GSM EFR coder. In comparison with standard GSM Full Rate coder, here the proposed coder offers other benefit of saving of 0.8kbps which can be useful for better error protection during channel coding while also provides satisfactory results for the parameters of Objective analysis.

In this paper, Gait Energy Image (GEI) has constructed to apply Principal Component Analysis (PCA) with and without Radon Transform (RT). The Radon Transform is used to detect features within an image and PCA is used to reduce dimension of the images without much loss of information. The side view of slow walk; fast walk and carrying a ball walk have been selected from the CMU MoBo database for experimental purposes. The two techniques achieved equal error rates (EER) of 94.23%, 82.28%, and 90.38% for PCA only and 96.15%, 82.70% and 92.30% for PCA with RT for slow walk, fast walk and carrying a ball walk respectively.

Optical character recognition is one of the most active branches of pattern recognition deals with different aspects of automatic recognition of written patterns. Among numerous techniques, systems, and software reported in the literature, Persian printed digits classification has not been attended a lot. In this paper, a consistent system for transformation-independent recognition of Persian printed numerals based on Hu moment invariants, which are invariant to translation, rotation, and scale has been introduced. Since utilization of these invariants tackles with some important issues such as noise sensitivity, compactness and invariance to reversal patterns, some operations to compensate these drawbacks have been done. In addition, a robust classifier named fuzzy min-max neural network has been used to encounter such a compact and overlapped feature space. Set of different experiments has been done and results show the proposed system is so successful to invariant classification of Persian printed digits.

The UBIRIS.v2 dataset is a set of noisy colour iris images designed to simulate visible wavelength iris acquisition at-a-distance and on-the-move. This paper presents an examination of some of the characteristics that can impact the performance of iris recognition in the UBIRIS.v2 dataset. This dataset consists of iris images in the visible wavelength and was designed to be noisy. The quality and characteristics of these images are surveyed by examining seven different channels of information extracted from them: red, green, blue, intensity, value, lightness, and luminance. We present new quality metrics to assess the image characteristics with regard to focus, entropy, reflections, pupil constriction and pupillary boundary contrast. The results clearly suggest the existence of different characteristics for these channels and could be exploited for use in the design and evaluation of iris recognition systems.

Spectral clustering is a powerful technique in clustering specially when the structure of data is not linear and classical clustering methods lead to fail. In this paper, we propose a spectral clustering algorithm with a feature selection schema based on extracted features of Kernel PCA. In the proposed algorithm, selecting appropriate vectors is dependent upon entropy of clusters on these vectors and weighting method is influenced by sum of the existence gap between clusters and entropy of the vectors. Tuning the parameters has a great effect on the results of spectral clustering techniques. In the ideal case, comparing our method with NJW and Kernel K-Means indicate the successful of the proposed algorithm.