i-manager's Journal on Pattern Recognition (JPR)


Volume 3 Issue 1 March - May 2016

Research Paper

Multilingual Speech Processing through MFCCs feature extraction formultilingual speaker identification system

Vinay Kumar Jain* , Neeta Tripathi**
* Research Scholar, Chhattisgarh Swami Vivekanand Technical University, Bhilai, India.
** Principal, Shri Shankaracharya Institute of Technology and Management, Bhilai, India.
Jain, V. K., and Tripathi, N. (2016). Multilingual Speech Processing through MFCCs feature extraction for multilingual speaker identification system. i-manager’s Journal on Pattern Recognition, 3(1), 1-6. https://doi.org/10.26634/jpr.3.1.8102

Abstract

The speaker identification systems work only in a single language environment using sufficient data. Many countries including India are multilingual and hence the effect of multiple languages on a speaker identification system needs to be investigated. Speaker identification system shows poor performance when training is done in one language and the testing in another language. This is a major problem in multilingual speaker identification system. The main objective of this research work is to observe the impact of the languages on multilingual speaker identification system and identifying the variation of MFCC feature vector values in multilingual environments, which will help to design multilingual speaker identification system. The present paper explores the experimental result carried out on collected database of multilingual speakers of three Indian languages. The speech database consists of speech data recorded from 100 speakers including male and female. The Mel Frequency Cepstral Coefficients (MFCC) as a front end feature vectors are extracted from the speech signals. The minimum, maximum and mean values of the feature vectors have been calculated for the analysis. It is observed that Rajasthani language has the larger values as compared to Hindi language and Marathi Language in minimum values of the feature vectors, where as Marathi Language has the larger values as compared to Hindi language and Rajasthani language in maximum values of feature vectors. The impact of the languages on multilingual speaker identification system has been evaluated.

Research Paper

Feature Diminish Based Nonlinear Support Vector Machine For Micro Classification Of Digital Mammogram Images

R. Manoharan* , R. Kalaimagal**
* Research Scholar, M.S University, Tirunelveli, Tamil Nadu, India.
** Research Supervisor, Government Arts College for Men, Nandanam, Chennai, Tamil Nadu, India.
Manoharan, R., and Kalaimagal, R. (2016). Feature Diminish Based Nonlinear Support Vector Machine for Micro Classification of Digital Mammogram Images. i-manager’s Journal on Pattern Recognition, 3(1), 7-15. https://doi.org/10.26634/jpr.3.1.8103

Abstract

The interpretation and analysis of medical images represent an important and exciting part of computer vision and pattern recognition. Developing a computer-aided diagnosis system for cancer diseases, such as breast cancer, to assist physicians in hospitals is becoming of high importance and priority for many researchers and clinical centers. It is a complex process to develop a computer vision system to perform such tasks. Breast cancer is the cause of the most common cancer death in women. X-ray mammography is widely used to screen women with an increased risk of breast cancer. Computer Aided Detection (CAD) systems have been developed to boost efficiency and accuracy in diagnosing cancer. This research presents the design of CAD for cancerous micro calcification classification in digital mammogram images based on Discrete Shearlet Transform (DST) and Kernel Principal Component Analysis (KPCA). The purpose of the Kernel Principle Component Analysis improved the classification accuracy by reducing the number of features. The implementation of Mammogram breast cancer detection is done using MIAS Database and MATLAB Tool. Results are shown in DST based NLSVM superior to the other conventional classifier techniques.

Research Paper

Cost Control Model of Power Grid Maintenance using Fuzzy Pattern Recognition Theory

M. Bhargavi* , S. Vijayalakshmi**
*-** Assistant Professor, Department of Computer Science and Engineering, Sree Vidyanikethan Engineering College, Tirupati, India.
Bhargavi, M., and Vijayalakshmi, S. (2016). Cost Control Model of Power Grid Maintenance using Fuzzy Pattern Recognition Theory. i-manager’s Journal on Pattern Recognition, 3(1), 16-22. https://doi.org/10.26634/jpr.3.1.8104

Abstract

In power enterprises, the construction process is complicated using lines and equipment maintenance, the cost is affected by meteorological and geographical factors, which influence mode in uncertain. In this paper, the authors use a predictive control model to control the project cost using fuzzy clustering method and the threshold intervals of the objective function in clusters. This model uses relative fuzzy operator to build fuzzy matrix, construct correlation between factors, and describe the factors' effect. Extract the cluster's Eigen function, define the boundaries of various clusters, and determine the type of the predicted points and the range of the objective function. When the actual cost of the maintenance project is within the range calculated by the cost model, then it is normal. If the actual cost exceeds this range, then further analysis of all the aspects of the cost is needed to find out the reason.

Research Paper

Image Steganalysis: Segmenting Stego Image using Watershed Method and Feature Extraction by MRF Method

B. Yamini* , R. Sabitha**
* Research Scholar, Department of Computer Science and Engineering, Sathyabama University, Chennai, Tamil Nadu, India.
** Professor and Head, Department of Information Technology, Jeppiaar Engineering College, Chennai, Tamil Nadu, India.
Yamini, B., and Sabitha, R. (2016). Image Steganalysis: Segmenting Stego Image using Watershed Method and Feature Extraction by MRF Method. i-manager’s Journal on Pattern Recognition, 3(1), 23-30. https://doi.org/10.26634/jpr.3.1.8105

Abstract

Transmitting the secret information by Steganography plays a vital role in Human Visual System (HVS). The carrier media such as image or audio or video can be used to hide the information. Steganalysis is a technique used to get rid of cheating, by identifying the hidden information from the carrier media. The identification of embedded information or message from the carrier media produces higher success rate to steganography methods. Image Steganography is the art of hiding the message or a file or an image by taking the image as carrier media. Based on the adaptable regions, the content is hidden and this method is termed as Adaptive image steganography. Dealing with retrieval of embedded content from the adaptable region of cover image is known to be Adaptive Image Steganalysis. The Blind Steganalysis is the ability to attack the stego image without the knowledge about steganography. Its Counter method, attacks the stego image by significant method used for steganography. In existing method, Enhanced canny edge detector is used to extract the features of the image better than other edge detectors, but smoothens the boundaries including noise and fails to identify the false edges. In the proposed method, Watershed method is used to segment the Adaptive regions from the stego image. The Markov Random Fields (MRF) extracts the features from the segmented adaptive region. The precision and recall is calculated after identifying the adaptive region with its payload location and hidden content using SVM (Support Vector Machine) classifier. An SVM is a binary classifier, classifies data by finding the best hyperplane which separates all data points of one class from the other class. After the classification and identification of payload location, the message is extracted from the hidden region by reversible two LSB (Least Significant Bit) bits.

Research Paper

Digital Image Watermarking Based On Gradient Direction Quantization And Denoising Using Different Techiniques

I. Kullayamma* , Sathyanarayana**
* Assistant Professor, Department of Electronics and Communication Engineering, SV University, Tirupati, India.
** Professor, Department of Electronics and Communication Engineering, Annamacharya Institute of Technology and Sciences, Tirupati, India.
Kullayamma, I., and Sathyanarayana (2016). Digital Image Watermarking Based On Gradient Direction Quantization And Denoising Using Different Techiniques. i-manager’s Journal on Pattern Recognition, 3(1), 31-42. https://doi.org/10.26634/jpr.3.1.8106

Abstract

Digital watermarking is the act of hiding a message related to a digital signal (i.e. an image, song, video) within the signal itself. In recent years, the phenomenal growth of the internet has highlighted the need for a mechanism to protect ownership of digital media. Digital watermarking is a solution to the problem of copyright protection and authentication of multimedia data while working in a networked environment. The authors propose a robust quantization-based image watermarking method, called the Gradient Direction Water Marking (GDWM), and based on the uniform quantization of the direction of gradient vectors. In GDWM, the watermark bits are embedded by quantizing the angles of significant gradient vectors at multiple wavelet scales. The proposed method has the following advantages: 1) Increased invisibility of the embedded watermark, 2) Robustness to amplitude scaling attacks, and 3) Increased watermarking capacity. To quantize the gradient direction, the DWT coefficients are modified based on the derived relationship between the changes in coefficients and the change in the gradient direction. In this paper, they propose fourdifferent denoising techniques for checking of the watermarking efficiency [15]. In various noise scenarios, the performance of the proposed denoised methods are compared in terms of PSNR and Correlation Coefficient. The Contourlet transform provides better PSNR when compared to other filter methods. The Correlation Coefficient observed that the Contourlet transform provides almost near to 1 which is ideal.